187
MYERS-BRIGGS TYPE INDICATOR PREFERENCE SCORES AND AGE By RICHARD IVOR KAINZ A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1985

MYERS-BRIGGS TYPE INDICATOR PREFERENCE SCORES

Embed Size (px)

Citation preview

MYERS-BRIGGS TYPE INDICATOR PREFERENCE SCORES AND AGE

By

RICHARD IVOR KAINZ

A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

1985

ACKNOWLEDGEMENTS

I would like to acknowledge, first, the patient support

and encouragement I have received from the members of my

supervisory committee. Particularly I would like to thank

my committee chair, Dr. Benjamin Barger, who was always

available when advice or review of the materials was

needed and Dr. Margaret K. Morgan, whose editorial assist­

ance went far beyond the call of duty. I owe a very deep

debt to Dr. Mary H. McCaulley, who during my apprenticeship

in psychology has been the greatest influence on my work: as

a mentor, as a model of academic and professional integrity,

and as a friend. Her dedication to humane ideals has been a

constant inspiration to my own efforts to be of service.

wi thout her guidance and support this work would not have

been begun, much less completed.

I also acknowledge the support of the Center for

Applications of Psychological Type (CAPT), which has allowed

me to use the CAPT Myers-Briggs Type Indicator data bank in

this investigation and has permitted me to take leave from

my duties when this project required my time. Glenn

Grennade, data processing manager for CAPT, contributed

greatly in preparing the data selection and item analyses

computer programs I have used in the study. I appreciate

the interest and support for this proj ect of the entire

staff of CAPT.

ii

I am thankful to the friends and family who share my

life and home for supportively accepting in their midst a

person writing a doctoral dissertation. Their cheerful

affection during the preparation of this manuscript is

greater evidence of their love than I could wish for. In

particular, Deidre Bryan provided invaluable assisstance in

the editing of the final manuscript. Most important has

been Maria Cristina's patience and loving support while this

work took up so much of the time and attention that were

rightfully hers.

Finally, I lovingly dedicate this work to the person

who taught me to love scholarship and who was my first

teacher, Maxine Lenore Kitterman, my mother. Her contribu­

tion is incalculable.

iii

TABLE OF CONTENTS

ACKNOWLEDGMENTS • ..

• • • • • • 11

TABLE OF CONTENTS . . iv

LIST OF TABLES · . . . . vi

LIST OF FIGURES . · . . ix

ABSTRACT x

CHAPTERS

I

II

III

IV

INTRODUCTION 1

Purpose of the Study • • • • . . . . . . . . • . . 1 Overview of the Problem . . . . . . . . . . . . . 3 Jung's Theory of Psychological Types ••..... 11 Description and History of the Myers-Briggs

Type Indicator . . . • . .. ..•.•... 28 History of the MBTI Item Weights • . . • • . . . • 33 Review of the Research Literature: Psychological

Type and Age • • . . . . . . . . . . . . . . . . 39 Summary of the Problem . . . . . . . . . . . . . . 49 Research Predictions • • • • . . • . • • • 57

METHODS • 68

Introduction . .....•• • • . . . 68 Selection of the Sample. • • • • . • . . • . • . • 68 Potential Sources of Sampling Bias • • . . . . . . 70 Analyses of Variance and Regression Analyses . . . 71 Homogeneity of Variance in the Sample . . • . 73 Analyses of the Item Weights . . . . . . . . . . . 73 Computer Software Used in Sample Selection and

Data Analyses . . . . . . • . . . • • . 76

RESULTS • • • 78

Introduction . . . . . • . . . . . . . . • . . . . 78 Analyses of Variance and Regression Analyses . . . 79 Analyses of the Item Weights . . .. ..... 98

SUMMARY, DISCUSSION, AND CONCLUSIONS.

Summary of the Study • . Discussion . . . • • . • . Conclusions . • • • • • •

iv

· •. 141

· • . • . 141 · • . . . 155 · . . . . 161

Recommendations for Further Investigation

APPENDIX: RESULTS FROM THE PILOT STUDY

REFERENCES

BIOGRAPHICAL SKETCH .

v

.162

. • . . .165

.170

.173

LIST OF TABLES

Table 1. Dominant and Auxiliary Attitude-Functions of MBTI Types . . . . . . . . . . . . . . . . . 63

Table 2. Predicted Changes of MBTI Preference Scores by Types. . . . . . . . . . . . . . •• 64

Table 3. MBTI Types Jung Grouped as Rational and Irrational. . . . .. ...... .... 65

Table 4. MBTI Types Sharing Dominant or Auxiliary Atitude-Fuctions • • • • . • • • . . . 65

Table 5. Predicted Changes of MBTI Preference Scores by Groups of Types . . . . . . . . . . • . . . 66

Table 6. Sample Frequencies per Gender for Age Groups. · .77

Table 7. Analyses of Variance: EI by Age Groups. · · · · 110

Table 8. Analyses of Variance: EI by Developmental Stages. . . . . . · · · · · 110

Table 9. Regressions of EI by Age: All Ages. · · · · 110

Table 10. Regressions of EI by Age: Stage Two. · · · · · III

Table II. Regressions of EI by Age: Stage Three. · · III

Table 12. Analyses of Variance: JP by Age Groups. · · · .114

Table 13. Analyses of Variance: JP by Developmental Stages. . . . . · · · · · .114

Table 14. Regressions of JP by Age: All Ages. · · · · · .115

Table 15. Regressions of JP by Age: Stage Two. · · · · · 115

Table 16. Regressions of JP by Age: Stage Three. · · 116

Table 17. Analyses of Variance: SN by Age Groups. · · · .119

Table 18. Analyses of Variance: SN by Developmental Stages. . . . . · · · · · .119

Table 19. Regressions of SN by Age: All Ages. · · · · · .120

Table 20. Regressions of SN by Age: Stage Two. · · · · · 120

vi

Table 2l. Regressions of SN by Age: stage Three. · · · · 121

Table 22. Analyses of Variance: Auxiliary SN by Age Groups. . . . . . · · . . · · · · · · · .124

Table 23. Analyses of Variance: Auxiliary SN by Developmental stages. · · · · · · · · · .124

Table 24. Regressions of Auxiliary SN by Age: All Ages ................... 125

Table 25. Regressions of Auxiliary SN by Age: stage Two ...••.•...•..•.•.. 125

Table 26. Regressions of Auxiliary SN by Age: stage Three .....••.•.•..•... 126

Table 27. Analyses of Variance: TF by Age Groups ..... 129

Table 28. Analyses of Variance: TF by Developmental stages ...

Table 29. Regressions of TF by Age: All Ages.

Table 30. Regressions of TF by Age: stage Two.

· .

Table 3l. Regressions of TF by Age: stage Three.

Table 32. Analyses of Variance: Auxiliary TF by Age Groups. . . . . . · · . . · · ·

Table 33. Analyses of Variance: Auxiliary TF by

· .129

· · · · .130

· · · · 130

· · · · 131

· · · · .134

Developmental Stages. •. ........134

Table 34. Regressions of Auxiliary TF by Age: Age Groups .........•........ 135

Table 35. Regressions of Auxiliary TF by Age: stage Two ..••.••........•.. 135

Table 36. Regressions of Auxiliary TF by Age: stage Three. . . . . • . .. .•.... 136

Table 40. Point Deviations By Age Groups: EI Items. .137

Table 41. Point Deviations By Age Groups: SN Items. .138

Table 42. Point Deviations By Age Groups: TF Items .... 139

Table 43. Point Deviations By Age Groups: JP Items. .140

vii

Table 44. Number of MBTI Items Having out-of-range Response Weights .•............. 164

Table 45. Summary Anovas By Age Groups: pilot Study ..• 165

viii

LIST OF FIGURES

Figure l. Mean EI Preference Scores by Age Groups. · · · 109

Figure 2 . Mean JP Preference Scores by Age Groups: Extraverts. · · · · · · · · · · · . . · · · .112

Figure 3 . Mean JP Preference Scores by Age Groups: Introverts. · · · · · · · · · · · · · · 113

Figure 4. Mean SN Preference Scores by Age Groups: Extraverts. · · · · · · · · · · · . . · · · .117

Figure 5. Mean SN Preference Scores by Age Groups: Introverts. · · · · · · · · · · · · · · .118

Figure 6. Mean Auxiliary SN Preference Scores by Age Groups: Introverts. . . . •. . ... 122

Figure 7. Mean Auxiliary SN Preference Scores by Age Groups: Extroverts. • . • .. . ... 123

Figure 8. Mean TF Preference Scores by Age Groups: Extraverts .•.....•..•....... 127

Figure 9. Mean TF Preference Scores by Age Groups: Introverts. . . . . .. . ..•..... 128

Figure 10. Mean Auxiliary TF Preference Scores by Age Groups: Introverts. . . . . . · · · 132

Figure 1l. Mean Auxiliary TF Preference Scores by Age Groups: Extraverts. . . . . . · 133

Figure 12. Mean EI Preference Scores: pilot Study. · · · 166

Figure 13. Mean SN Preference Scores: pilot Study. · · · 167

Figure 14. Mean TF Preference Scores: pilot Study. · · · 168

Figure 14. Mean JP Preference Scores: pilot Study. · · · 169

ix

Abstract of Dissertation Submitted to the Graduate School of the University of Florida in Partial Fulfillment of the

Requirements for the Degree Doctor of Philosophy

MYERS-BRIGGS TYPE INDICATOR PREFERENCE SCORES AND AGE

By

Richard Ivor Kainz

December, 1985

Chairman: Benjamin Barger Major Department: Clinical Psychology

The researcher analyzed 21,696 randomly selected Myers-

Briggs Type Indicator (MBTI) records (a) for differences of

mean MBTI scores between eight age groups (IS-17, 18-20, 21-

24, 2S-29, 30-39, 40-49, SO-S9, 60 and older), and two

developmental stages (adolescence to mid-life [IS-39], mid-

life and beyond [40 and older]); (b) for predicted patterns

among association trends between MBTI scores and age within

two developmental stages; and (c) for differences in the

scoring weights of MBTI items for appropriateness among

differently aged respondents.

The MBTI questionnaire (designed to assess the

preferences for personality orientations Jung described in

his theory of psychological types) yields scores on four

x

indices: Extraversion-Introversion (EI), Sensing-Intuition

(SN), Thinking-Feeling (TF), and Judgment-Perception (JP).

means

MBTI score Investigators have found differences among

suggesting trends of association with age unanti­

theoretically cipated in theoretical discussions and

inconsistent. However, previous investigators failed to

include equal numbers of respondents having opposite

preferences and did not maintain Jung's distinction between

dominant and auxiliary personality orientations. These

analyses correct those errors.

Response weights for eight age groups were examined

separately by group, replicating Myers's procedures for

assigning weights. Associations of items with age group

were tested with supplementary chi-square analyses.

Small mean differences and trends of association

suggest that extraversion and perception decrease, and

sensing, intuition, and judgment increase with age. Only

Sensing-Intuition analyses support predictions.

Item analyses indicate Myers's procedures would now

change weights for one third (for males) to one half (for

females) of the items. Response weight changes varied by

age groups. Chi-square analyses suggest response and age

associations for 81% (for females) to 85% (for males) of the

items.

These analyses of MBTI preference scores and age

differences are confounded by response weight differences

xi

for differently aged respondents. Reliable analyses cannot

be performed until the response weights are updated.

The investigator recommends updating the MBTI response

weights, developing normative MBTI data for older

respondents, and distinguishing between generational and

longitudinal effects in future research.

xii

CHAPTER I INTRODUCTION

Purpose of the study

Carl Jung, the Swiss physician and psychologist,

published a theory of psychological types in 1921. within

this theory, Jung described and organized his observations

of personality differences that he believed to be basic and

differing orientations of normal consciousness. At its core

the Jungian typology is a cognitive rather than motivational

theory, based on Jung's observation that conscious

information processing and decision making are organized by

two complementary but opposing functions of perception

(sensation and intuition), and by two complementary but

opposing functions of judgment (thinking and feeling). Jung

observed also that the functions operate in one of two

complementary but opposing general orientations, or

atti tudes, which differentially focus consciousness toward

either external objects (extraversion) or internal objects

(introversion). The habitual use of one of these attitude-

function combinations, to the relative exclusion of the

others, results in the development of characteristics of

personality that Jung considered to be a psychological type.

Jung considered the development of habitual preferences for

a dominant attitude-function combination allied with a

lesser developed auxiliary attitude-function combination to

1

2

be an enduring and stable mark of normal ego growth (Jung

1968, 1971; Jacobi, 1973).

Katherine Briggs and Isabel Briggs Myers developed the

Myers-Briggs Type Indicator (MBTI) (Myers, 1975, 1977) to

make operational Jung's theory and published the MBTI Form F

in 1962. After 14 years of use as a research instrument,

the MBTI became widely available in 1976 for applied use.

The MBTI is now used with increasing frequency in a variety

of circumstances, including individual, marital, family,

pastoral, academic, and career counseling, and also in

management or organizational development consultation for a

variety of public and private institutions. Although

investigators focused research and applications of the MBTI

between 1962 and 1976 on samples of high school and college

students, broader applications are generalizing the

successful research reported for student samples to adult

populations. Respondents who comprised the original samples

used in the development of the MBTI were mature adults.

However, in a series of efforts to restandardize the MBTI

respondents were students from the fourth grade into the

first year of college. The restandardized scoring proced­

ures are now applied for respondents of all ages.

The purpose of this study was to analyze data from a

large and previously unavailable sample of MBTI records to

see whether differences predicted from theory for the

relative importance of psychological type preferences for

3

different age groups are reflected in the MBTI scores. If

differences in preference scores can be demonstrated for age

groups, the results of this study could lead to refinements

of the scoring procedures and/or interpretation of the MBTI

reports.

overview of the Problem

The Myers-Briggs Type Indicator (MBTI), Form F (Myers,

1975, 1977) is a self-report questionnaire explicitly

designed to assess the psychological type preferences Jung

described by in his theory. The questionnaire consists of

95 forced-choice items scored on four scales and 71

additional items included for research to individualize MBTI

reports. Scoring of the four scales is used to indicate a

respondent's preferences on the three dichotomous

personality orientations Jung postulated and a fourth

dichotomous personality orientation that Briggs and Myers

considered implicit in Jung's theory and necessary to

properly interpret relationships among orientations ex­

plicitly described by Jung.

Scale scores are considered to be an intermediate step

to the sorting of respondents into personality type classi­

fications; however, the authors of the MBTI and numerous

other researchers have employed the scale scores directly in

the study of the relationship of the typology to a variety

of criterion variables. Investigators have studied major

demographic variables of gender and occupation with respect

4

to the MBTI scales. Counselors often consider the

respondent's age in the interpretation of the MBTI, but have

little explicit theory and no reliable research to guide

their interpretative adjustments.

Scoring of the MBTI results in a preference score

(one for each of the four indices) expressed as a letter

indicating the direction of the preference and a number

quantifying the strength of the preference. The letter

indices are combined to produce a type formula representing

the individual's psychological type. The four scales are

Extraversion-Introversion (EI) , Sensing-Intuition (SN) ,

Thinking-Feeling (TF), and Judging-Perceptive (JP). A type

formula combines one of each of the pairs, such as ENFP or

ISTJ. Each respondent is classified into one of the 16

possible combinations of the preferences and is assigned

four preference scores, such as E17, N33, F25, P13, or lIS,

S23, T29, J9. In these two examples the individuals would

differ on all four preferences. The 16 type combinations

are shown here in the matrix form known as a type table,

usually used to display the frequencies of types in a

sample.

I ISTJ I ISFJ I INFJ I INTJ I

I ISTP I ISFP I INFP I INTP I

I ESTP I ESFP I ENFP I ENTP I

I ESTJ I ESFJ I ENFJ I ENTJ I

5

Broad searches of the computer databases of the

psychological research literature disclosed only one study

(Bloch, 1978) of the relationship of MBTI scale scores to

the respondents' age. Bloch found that regression analysis

associated higher scores for three (sensing, intuition, and

judging) of the eight preferences with increasing age. An

unpublished student paper submitted to the Center for

Applications of Psychological Type for inclusion in the

Isabel Myers Memorial Library collection of research

materials also reported MBTI types at two age levels. This

student, Susan Schaeffer (1974), found extraverts to occur

more frequently among a small sample of students over the

age of 35 ; however, she did not use scale scores in her

analyses.

Due to paucity of available data on MBTI scale scores

and age, the literature search also included studies of the

Jungian Type Survey (JTS) (Gray & Wheelwright, 1946). The

JTS (11th ed.), is a 75-item self-report questionnaire

scored on three scales representing the three dichotomous

personality orientations of Jung's theory. As with the

MBTI, the intent was to classify respondents into the

psychological types described by Jung. stricker and Ross

(reported in Myers, 1975) obtained high correlations among

47 male college students for the three scales common to the

two instruments. Myers concluded that after correction for

attenuation the three scales have their entire true-variance

6

components in common. Thus, JTS data should correlate with

MBTI data.

Al though investigators have done less research with

the JTS than with the MBTI, two have related the JTS scales

to the age of the respondents. Gray (1947b, 1948), using

regression analyses, found all three of the JTS scales to be

associated with age (introversion, sensing, and thinking

increased) in samples ranging over six decades. Driver

(1975) in a study of the elderly (mean age, 76.5; range, 55-

102 years) reported an increasing score with the progression

of age only for introversion.

In an analysis of the results reported for the MBTI

and the JTS altogether, investiigators twice reported

significant increases for introversion, once for

extraversion, twice for sensing, never for intuition, once

for thinking, never for feeling, once for judging, and never

for perceptive. The judging and perceptive preferences are

not assessed directly on the JTS scales.

Gray (1948) also reported the JTS scales to be

associated with gender (males having higher scores for

introversion, sensing, and thinking). He did not report

analyses in which he considered gender and age simul tan­

eously.

Prior to conducting the literature search summarized

above, I conducted an exploratory analysis of 52,848 MBTI

Form F records contained in the MBTI data bank maintained by

7

the Center for Applications of Psychological Type. This

sample included all respondents reporting an age of 15 years

or greater (excluding scores for respondents omitting 35 or

more items). These MBTI records had been submitted to CAPT

for computer scoring from 1978 through 1982.

separate one-way analyses of variance on

preference scores at eight levels of age for

I performed

the eight

each sex.

These analyses produced significant differences among the

means of the eight preference scores when grouped by age

levels (see Appendix A, Table 42). However, the results

differ with respect to gender. All analyses of data from

females resulted in significant differences; two of the

analyses of data from males (extraversion and feeling) fell

just short of the .05 probability level. An inspection of

the means (plotted in Figures 12 through 15 in the Appendix)

suggests that more carefully controlled analyses are likely

to reveal systematic trends in the preference scores for

changes in age differentially for the respondents' gender

and type classification. A gender-by-age interaction

appears most probable for the extraverted and feeling

preferences.

These exploratory analyses are the first in which an

investigator attempted to separate the influence of gender

and age on scale scores of Jungian type indicators and in

this they are an improvement over the earlier studies. Yet

they are methodologically deficient in several respects,

8

with some problems in common withthe earlier studies. In

all of the studies summarized here the frequencies of the

ages, gender (except Gray, 1947b, 1948), and psychological

types of the respondents varied widely within each sample,

allowing the exaggerated influence of the over-represented

gender, ages, and types to be included in the analyses.

Only Bloch (1978) and Driver (1975) tested assumptions of

the comparability of variance of the dependent measures.

The findings of the published research and exploratory

analyses are of theoretical importance in so far as they

imply that individuals may change type preferences as they

age. A part of Jung' s theory is the idea that type

preferences are inborn or so thoroughly set early in child­

hood as to be a lifelong pattern for the individual's

psychological differentiation. within the premises of the

theory an extravert does not later become an introvert, or

vice versa. However, theory does not preclude an extra-

vert, for example, from developing a greater appreciation of

and skill with introversion. In fact, such development is

considered necessary to overcome the one-sided development

of youth, making possible higher levels of personality

integration in later life. Nevertheless, an assumption of

the theory is that the type preferences differentiated by

early development will hold their relative dominance

throughout life.

9

The developers of the MBTI undertook the construction

with careful attention to the correct positioning of the

cutting scores, or mid-points, of the scales used as indices

of type preferences (Myers, 1961; Myers, 1975; Myers &

McCaulley, 1985). Myers has never asserted that the

respondent's true type preferences that Jung described in

his theory will always be reflected in the MBTI report.

McCaulley (Myers & McCaulley, 1985) cites several

circumstances that may interfere with the respondent's

ability or motivation to report type preferences on MBTI

scales. Notable among these are various forms of distortion

to conform with parental, societal, and occupational demands

fel t by the respondent, or administration of the MBTI to

respondents whose personality development is poorly

differentiated or who are in crisis. Myers and McCaulley

advised counselors that the error of measurement should

always be considered in the reporting of MBTI scores to

respondents, particularly when the preference scores are

low. The need for sensitive interpretation is advised on

the basis of a general, rather than a systematic, source of

error of measurement.

As a practical matter, the classification of

individuals into their types according to their positions on

the scales may be subject to systematic error if the items

and/or their associated weightings have differential

sensitivity to the manifestation of the type preferences at

10

different ages. This could arise through the inclusion of

item responses which reflect too great a specificity for

age-appropriate behaviors, or through changing perceptions

of the social desirability of certain responses as a result

of greater maturity and integration. If the items and the

item weights are not tested across the range of ages for

which they are to be used, the true cutting score would

shift from that prescribed by the scoring procedures. The

result would be the misclassification of individual

respondent's type preferences. Error of measurement is

unavoidable in any instrument, but when it can be determined

to be systematic every reasonable effort should be made to

eliminate it.

I undertook the present study for three purposes: (a)

to formulate a set of research predictions which may (in a

manner consistent with Jung's theory of psychological types

and the elaboration of that theory by Myers and others)

account for relationships associating MBTI preference scores

with age; (b) to provide a data analysis employing proper

statistical controls to identify age and age-by-gender

interactions with the MBTI preference scores, in order to

test the hypothetical formulations; and (c) to examine the

MBTI items scored for type preferences for the presence of

differential weights for age or age-by-gender combinations,

using the same item-analysis procedures described by Myers

11

(1962, 1978) with the addition of controls for age

differences.

Jung's Theory of Psychological Types

In designing this study I was concerned with the

association of MBTI preference scores with respondents'

ages. To develop research predictions that may account for

such an association, one must understand the Jungian theory

that guided the construction of the MBTI. The following

discussion is an explanation of the theoretical base for the

research predictions to be proposed.

Jung's theory of psychological types, initially

published in 1921, presents, "the various aspects of

consciousness, the various attitudes the conscious mind

might take toward the world, and thus constitutes a

psychology of consciousness regarded from what might be

called a clinical angle" (Jung, 1961, p. 207). Jung

considered the theory to reflect the clinical angle because

it derives not from his previous theoretical formulations,

or from experimental studies (as with his word-association

research), but from both his casual and his systematic

direct observational experience of his patients, colleagues,

and associates. Jung clearly continued to employ the

typology throughout his years of medical practice because he

found it essential to his psychotherapeutic work: "one

always has to answer people in their main function,

otherwise no contact is established" (Jung, 1968, p. 157).

12

Nevertheless, after completing his work Psychological Types

(Jung, 1971), he did not explicitly return to the subject to

offer further theoretical elaboration.

A possible explanation for the lack of elaboration may

lie in his frustration with the reception which the work

received (perhaps more with the admirers than the critics),

as evidenced by the following remarks: "there is little

practical purpose in making the typology still more

complicated when not even the elements have been properly

understood" (Jung, 1971, p. xii) and "my typology is far

rather a critical apparatus serving to sort out and organize

the welter of empirical material, but not in any sense to

stick labels on people at first sight" (Jung, 1971, p. xiv).

Clearly Jung intended his theory of personality types to be

understood within the broader context of his psychological

views of human development, rather than as an isolated

system of classification. Edinger asserts "Jung made no

effort to present a systematic theory of psychological

development" (1968, p. 8), and his followers who have done

so, e.g., Neumann (1954), have not explicitly incorporated

the type theory. However, some prominent Jungian theorists

regard the type theory as the initial formulation of Jung's

core developmental concept of individuation. Meier (1971)

argues against the view that Jung had abandoned the type

theory, "for what he really did was to devote the rest of

his life to the amplification of the dynamics of the

13

individuation process, the blueprint of which is already

neatly given with the typology" (p. 279). Fordham (1972)

concurred and suggested that the failure of Jung's followers

to elaborate the developmental aspects of the type theory is

"because the essence of Jung' s thesis has been ignored and

the stable (aeternal) view of types has replaced the dynamic

one altogether" (p. 114). Jung first presented the concept

of individuation in Psychological Types, though only through

definition in the final chapter.

The body of his work clearly shows Jung's overriding

concern with the problems of psychological maturation beyond

mid-life. At this age the individual has successfully

maneuvered through the vicissitudes of childhood and youth

and is prepared to actively embark on the process of

individuation. Individuation, in Jung's theory, is the

process by which a person becomes psychologically

differentiated from the collective in a manner that is self-

aware or conscious. For Jung the collective refers to

psychic contents (e.g., ideas, opinions, values, and general

concepts) which belong not to one individual but to many.

The task of childhood and youth is to come to terms with and

to adapt to the collective, thereby establishing a position

in one's society.

At the same time, culture effects a differentiation of the function that already enjoys a better capacity for development through heredi ty. In one man it is the capacity for thought, in another feeling, which is particularly amenable to development, and therefore impelled by

14

cultural demands, he will concern himself in special degree to developing an aptitude to which he is already favorably disposed by nature. Its cUltivation does not mean the function in question has an a priori claim to any particularly proficiency, on the contrary, one might say it presupposes a certain delicacy, lability, pliability, on which account the highest individual value is not always to be sought or found in this function, but rather, perhaps is developed for a collective end.. The differentiated function procures for him the possibility of a collective existence, but not always that satisfaction and joie de vivre which development of the individual values alone can give. (Jung, 1971, p. 75)

In this schema, Jung probably differs more from

Freud than from Adler. Mosak and Dreikers (1972) present

the Adlerian view of ego development: "At one end of the

spectrum we have the primary sense of inferiority; at the

other the mutual interdependence of people. As a rule of

thumb, to the extent that an individual possesses social

interest his feelings of inferiority subside or are

eliminated" (p. 46). However, for Jung the adaptation to

social interest results in a one-sided and unbalanced

development of the personality through repression or

suppression of latent capacities, which do indeed subside

during the first half of life only to reemerge later. "The

overdifferentiation of the superior function, which is

almost inevitable with the passage of the years, results

nearly always in tensions that are among the main problems

of the second half of life" (Jacobi, 1973; p. 18). Yet,

Jung argues that this one-sidedness of development not only

is necessary for adapting to the needs of collective life,

15

but may even be the impetus for the development of

consciousness itself. He does not describe an adult without

one-sided development as admirable, healthy, or even normal,

but rather as primitive and undifferentiated. A critical

task of youth is to identify and develop a differentiated

and adapted function. Jacobi (1973) marks the end of

adolescence as the time at which the dominant function will

have attained clear ascendancy and association with "ego­

consciousness."

Jung's

psychological

description of the characteristics of the

types roughly corresponds to the level of

development at which the dominant attitude-function is most

exclusively associated with ego-consciousness. The

following summary presents the specific constructs of the

typology and the theoretical dynamics of the constructs.

The dynamic relationships postulated to exist among these

constructs provide the direction and necessity for further

psychological development.

Jung postulates two pairs of complementary

psychological functions and a pair of general attitudes by

which consciousness is oriented. The four functions are

labeled sensing, intuition, thinking, and feeling; the two

general attitudes are extraversion and introversion. The

first pair, sensing and intuition, are functions of

perception, being the processes by which information comes

into ego-consciousness.

16

Sensing is confined to the data that come to the

physical senses and represent immediate, present reality.

As a function of consciousness, it is not simply a passive

reception of sensory stimuli, but an active identification

or recognition of specific objects in consciousness. Jung

describes differentiated sensation as a conscious function

in that its operation is consciously directed and focused.

That is, the sensing function does not just see, it looks;

does not just hear, but listens. In the act of such looking

and listening, more details of the separate elements of

experience are brought into consciousness.

Intuition, by contrast, is vaguely defined as a

perceptual function of consciousness whose operation is

unconscious and whose contents appear in consciousness whole

and complete. The contents of intuition are perceptions of

relationship that go beyond what is directly available to

the senses. Marshall (1968) extends the definition,

describing intuition as a function whose operation results

in a restructuring of the original separate elements of

experience into a complete whole. In its most basic

manifestation, intuition is observed in the capacity to

recognize geometric figures in the relationship among

separate dots, Le., to perceive gestalts; in its

differentiated form it can be the finely tuned "third ear"

of the experienced psychotherapist. By completing the whole

from the elements, intuition perceives possibilities that go

17

beyond the specific information present, to a perception of

a possible reality.

Sensing and intuition are defined as opposing and

complementary functions that cannot operate simultaneously.

Marshall (1968) helped to make this clear in contrasting the

analytic quality of sensing with the synthetic quality of

intuition:

Thus, restructuring of experience may occur in either an analytic or a synthetic direction. One's perception may get nearer the elements or nearer the organizing principles. Either direction involves both a gain and a loss of information. Analysis reveals more detail about a smaller area; synthesis reveals more general but less detailed patterns. (p. 20)

The second pair of functions, thinking and feeling,

similarly contrasts alternative conscious acts of judgment

by which the objects of perception are considered for their

merit and rational conclusions are drawn. The thinking

function judges the relationship among objects according to

laws, concepts, rules, paradigms, principles, standards, or

any other method of relationship that may be separated from

evaluative associations, and finds the relationship valid or

invalid. The feeling function determines the relative value

among objects along a dimension of acceptance or rejection.

Exercising the feeling function makes it possible to

appreciate the importance of an object or the importance of

the relationship among objects. For example, thinking may

lead a man to the judgment that, as a gift, an electronic

calculator would provide his wife with extra hours of

18

leisure (assuming

feeling would lead

she handles the family finances),

him to celebrate Valentine's Day

but

by

giving her a dozen red roses.

In the infant, according

functions exists in a passive,

to Jung, each of

undifferentiated,

the

and

unadapted state. Through interaction with the environment

one function becomes increasingly more active due to its

greater capacity to adapt to the collective. Through its

activity this leading or dominant function separates from

the other functions and initiates a relationship of object

and subject, with respect both to external objects, of the

environment and to internal objects as states of the

organism. When the orientation of the function directs its

activity toward external objects, it takes on the attitude

that Jung defined as extraversion. When it orients toward

subjective states as objects, the function takes on the

attitude Jung defined as introversion. Ego, or the ego­

complex, arises as the subject of the object-subject

relationship, and consciousness is the product of the

operation of the active differentiated function.

The effective adaptation of the most active, or

dominant, function requires the inactivity of the less

adapted functions insofar as they interfere with the

activity of the dominant function.

effectively until it is separated

Sensing cannot operate

from intuition. The

failure of differentiation results in perception of objects

19

that do not remain stable, being fused with their context or

surroundings; intuition that is not differentiated from

sensing produces the perception of relationships which

quickly fragment into unrelated specificity. Similarly,

thinking cannot become reliable in discerning valid

relationships among objects while feeling is actively

relating the value of the object to itself as the subject.

The most adaptive function takes its priority (as the

activity associated with the ego) through the repression or

exclusion from consciousness of the operations of the less

differentiated functions.

The extreme dependency of the child upon family and

society makes adaptation to the collective into which it is

born an imperative task of childhood. The dominant function

is most amenable to adaptation to the collective; this leads

to the most successful balancing of object-subject

conflicts. The adaptive success of the dominant function

establishes a recurring sequence wherein its preferred

activity is more frequently exercised and further

differentiated. Through exercise, the dominant function

increases its usefulness as a tool for adaptation and its

increased usefulness ensures its position as the preferred

activity of consciousness. This sequence accelerates and

stabilizes as persons near adulthood. Jung described

persons as psychological types when they orient their

20

conscious activity by use of their preferred function in a

habitual manner.

During the period of development when the

differentiation of the dominant function is progressing most

rapidly, the less preferred functions continue to develop.

But their operation is for the most part outside of

consciousness and is passive. They provide a necessary,

unconscious substrate for conscious activity. For example,

if intuition is the dominant function, its activity relates

sensory data which are passively perceived by sensing, have

been appreciated as relevant to the subject by feeling, and

whose relationship is judged valid by thinking.

Consciousness does not observe or direct the participation

of the passive functions, but receives the perception whole

and complete as a relationship or possibility present in

awareness. In another individual the active, dominant

function may be either sensing, thinking, or feeling. For

those individuals, as well, the dependency of the operation

of the function on the operations of the passive, sub­

ordinate functions is just as inescapable. While the

operations of the undifferentiated functions are "invisible"

to the individual, they are not so to observers who more

easily see the less mature and unadapted aspects than indi­

viduals permit themselves to acknowledge (Jung, 1971;

Jacobi, 1973; Myers, 1975).

21

The typology would be more obvious, far simpler,

easier to assess, and less useful if the process stopped

here. The adaptation by way of one function that takes on

one attitude would lead to a one-sided personality that has

narrow limitations. If the dominant function is one of

perception (8 or N), judgment remains passive and outside of

conscious control. Conversely, development of a judging

function (T or F) leaves perception passive and outside of

conscious control. Conscious activity is limited by the

emphasis of the dominant function. This one-sided

development of concsiousness is further limited because, in

Jung's model, the dominant function typically appears in one

of the two attitudes that orient consciousness to its

objects. If the dominant function takes the extraverted

attitude, only the individual's relations to external

objects are adapted and differentiated; internal objects are

recognized as belonging to the environment, often through

proj ection of the inner state onto an external obj ect. On

the other hand, an introverted dominant function does not

consciously relate to external objects as distinct from the

individual's own subjectivity, resulting in a vulnerability

to introjection. A mature and developed personality

requires balance in both perception and judgment, and in the

capacity to relate consciously to both external and internal

objects. The attempt of the immature personality to adapt

to the demands of the collective through the exclusive focus

22

on the dominant function, while ignoring the less adaptive

functions, inevitably begins to fail as the individual

reaches limits of the adaptability of the preferred

attitude-function combination. (Jung, 1971; Jacobi, 1973;

Myers, 1975).

In infancy, reliance on the dominant function provides

stability in emerging consciousness. The infant's failure

to exclude the opposing functions from its nascent,

minimally stable consciousness might precipitate the

dissolution of the ego-complex. However, once the

individual has attained a stable consciousness, the rigid

preference for the dominant and exclusion of the other

functions threatens the further growth of the personality.

When the adapted, conscious function is sufficiently secure,

it may permit the development of a second function of

conscious activity. This function will not be the most

ignored function, i.e., the function directly opposed to the

dominant; further, the second function will not have the

attitude of the dominant (Jung, 1971; Jacobi, 1973; Myers,

1975) .

The secondary function has a balancing effect, and

"always one whose nature is different from, though not

antagonistic to, the primary function" (Jung, 1971, p. 406).

For example, an introverted perceptive function may

associate with either of the extraverted judging functions,

or an extraverted perceptive function may associate with one

23

of the introverted judging functions. The development of the

secondary, auxiliary, function is generally overlooked in

what Fordham (1972) describes as the static interpretation

of the type theory. In the dynamic interpretation, with the

acceptance within the ego-complex of the development of an

auxiliary function comes the internalization of conflict as

an activity of consciousness. with his conceptualization of

differentiated ego functions having opposing aims and

autonomous operations Jung anticipated the contributions of

Hartmann to psychoanalytic theory.

In his seminal work Ego Psychology and the Problem of

Adaptation (1958), Hartmann introduced the concepts of

autonomous ego functions and the conflict-free ego sphere,

and adopted Kris's (1934) concept for relating the

functions, "regression in the service of the ego."

Hartmann specifically includes discussion of "perception,"

the "synthetic function," "goal-rational" action, "value-

rational" action, "objectively reality syntonic" action, and

"subjectively reality syntonic" action. These constructs

seem more familiar and less awkward if they are recognized

as Jung's sensing, intuition, thinking, feeling,

extraversion, and introversion, respectively. Hartmann

acknowledges the problem of the one-sidedness of the

developing personality, though his assessments appear to be

made from the perspective of an extraverted thinker:

If we keep in mind what we said above about synthesis and rank order [of the ego functions],

24

we arrive at a very different picture: the optimal role of goal-means thinking in adaptation is decided by the maturity, the strength, and the structure of the ego.

Why then is this picture so distorted? Because in it a particular ability has taken the place of all other mental functions. The picture becomes more human if we think of intelligence as organizing rather than taking the place of all other functions. At a certain level of development intelligence becomes aware of its own role as one function among others, sees its own activity in correct perspective among the other mental tendencies. (Hartmann, 1958, p. 69)

The admission of other functions into the conflict-free

ego sphere is accomplished by "regression in the service of

the ego." Thus, in Jungian terms the individual

accomplishes the development of the auxiliary function by

temporarily releasing the hold of the dominant function.

Jungians have generally ignored the dynamics and kept the

typology Jung employs to describe it. Freudians have

refused to assimilate the relativity of perspective implied

by alternative patterns of personality development because

that would undermine the scientific authority they claim for

their observational methods. Therefore, they have largely

ignored the implications of the typology but have readily

kept the dynamics.

At the end of adolescence, according to Jacobi's

(1973) summary of Jung's typology, the individual has

consolidated ego-consciousness, which is associated with the

dominant function; by the middle of life "all the functions

should fall into their proper order and be appropriately

25

differentiated" (p. 14). Though she does not explicitly

describe the development of personality from the end of

adolescence to the middle of life, the implication is clear

that this period must involve the growth and development of

the auxiliary function, as well as the further strengthening

of the dominant function. Experience with the internal

consistency of the MBTI scales would suggest that the

emergence of the auxiliary function may be reliably reported

by most subjects just past the age of puberty (Myers, 1962;

Myers & McCaulley, 1985). Though authors disagree on the

specific ages at which individuals will achieve a given

level of development, the period between middle adolescence

(about age 15) and mid-life (35-45 years) is probably the

greatest period of differentiation for the auxiliary

function.

with successful development, at mid-life the conscious

activity of the ego-complex should be marked by (a) the

dominance of a strongly differentiated function of either

perception or judgment using either the extraverted or

introverted attitude, and (b) the presence in consciousness

of an auxiliary function that is somewhat less differen­

tiated than the dominant. The auxiliary differs from the

dominant in two ways in being a complementary function (of

perception or of judgment), and in operating through the

complementary attitude (extraversion or introversion). In

its structural and dynamic development the ego-complex

26

attains the capacity to maintain stable obj ect relations

with both external and internal objects. The developed ego

has served its necessary function of adapting the individual

to the conflicts arising between the organism and its

environment, and providing conscious mechanisms for

internalized conflict management. For Jung, the yardstick

of the effectiveness of the adaptation of the ego-complex is

its relationship to the collective. The development of this

tool for adaption is the task of childhood and youth. At

mid-life the stage is set for the adaptation of

consciousness to the needs of individuality; the stability

of the ego-complex provides a perspective from which

consciousness can relate to the more deeply unconscious

aspects of the personality described by Jung (Jung, 1968,

1971; Jacobi, 1973).While the deep unconscious psychological

structures described by Jung are of considerable interest in

other contexts, for the purpose of understanding the

relationship of the type functions it is sufficient to

recognize two points:

(1) Consciousness in the process of individuation becomes

less closely identified with the ego-complex, and there­

fore less identified with the dominant and auxiliary

functions. The identification of consciousness becomes

associated with "self," which Jung understands to be the

totality of the personality; the self is partly knowable

27

and observable by consciousness and is partly unknowable

and transcendent (Jung, 1968; Jacobi, 1973).

(2) The functions that have not been differentiated and

integrated by the ego-complex may be further developed

in consciousness, especially the function opposite

the auxiliary. But this "inferior function" can

never become fully conscious, perhaps only

because life is finite and the task is too great for

a lifetime (Jung, 1968; Jacobi, 1973).

In summary, the theory of the development of (a) the

dominant, (b) the auxiliary, and (c) the less-preferred

functions throughout the life span forms the framework for

the analyses of the MBTI preference scores in this research.

In attempting to understand the expression of preferences

for type functions across the age span for which data are

available, differring expectations occur for two periods.

The first period, from approximately 15 to 40 years, should

show a continuing increase in the preference for the

dominant function, and an even greater increase in

preference for the auxiliary function. Associated with the

greater increase in preference for the auxiliary function

should be a trend for increased preference of the attitude

taken by the auxiliary function. This increased preference

for the auxiliary should result in a reduction of the

preference for the attitude of the dominant function. In

the second period, approximately age 40 and beyond, the

28

trend to increase the preference for the dominant function

should falloff, while the trend toward increasing

preference for the auxiliary should continue. In a later

section these expected trends will be stated with specific

reference to MBTI preference scores.

Description and History of the Myers-Briggs Type Indicator

The focus of this section is on the psychometric

features of the Myers-Briggs Type Indicator that are

relevant to the data analyses of this research. The Myers­

Briggs Type Indicator (also known as the MBTI, and the Type

Indicator) is a self-report questionnaire consisting of 95

forced-choice items scored on four scales: Extraversion­

Introversion (EI), Sensing-Intuition (SN), Thinking-Feeling

(TF), and Judgment-Perception (JP). The EI, SN, and TF

scales are explicitly constructed to assess the function and

attitude preferences described by Jung's typology.

Each item offers a choice of responses reflecting the

dichotomous index for which the item is scored. Current

scoring procedures provide separate weighting for each

response; thus a response may add two, one, or zero points

toward a point total for its respective preference. Scale

scores are computed by finding the difference of the point

totals for each pair. Since the primary purpose of the MBTI

is to indicate the Jungian preference types, and tied scores

would not indicate preferences, the point total differences

are doubled and one point is added or subtracted so that no

29

resulting preference score may be zero. The decision about

whether a point should be added or subtracted follows rules

devised by the authors of the MBTI. On the TF scale the

rule adds one point for males in the thinking direction or

one point for females in the feeling direction--by following

the rationale that the setting of item weights on that scale

may have overvalued evidence contrary to the prevailing

preference for each gender. On the remaining scales, one

point is added in the direction of the preferences occurring

less often in the general population--following the

rationale that evidence contrary to popular preferences is

to be believed. The resulting preference scores consist of

two parts, a letter indicating the direction of the

preference and a number indicating the "strength" of the

preference. The type is indicated by the four letters of

the preference scores.

Three indices indicate preferences for an attitude

(EI), a perceptive function (SN), and a jUdging function

(TF) . It is not possible to determine from these three

indices alone which preferred function (the perceptive or

the judging) is the dominant and which function is the

auxiliary. Briggs and Myers have resolved this ambiguity by

the addition of a fourth index labeled Judgment-Perception

(JP). In format and scoring procedures this scale parallels

the other three. Its item content is constructed to provide

for each item the choice between the use of (a) a judgment

30

function or (b) a perception function in relating to the

environment (implying orientation through extraversion). In

the interpretation of this scale Briggs and Myers have

exploited the dynamic relationships Jung postulated to exist

among the type functions. That is, for extraverts the

dominant function is extraverted and the auxiliary function

is introverted. For introverts the dominant function has

the introverted attitude, and the auxiliary, the extraverted

atitude. By indicating whether the respondent has a judging

or perceptive function associated with the extraverted

attitude, the JP scale leads to the identification of the

dominant function. If the JP scale indicates that a judging

function is extraverted, the dominant function for

extraverts is indicated by the TF scale. If the JP scale

indicates extraversion of a perceptive function, the

dominant function of an extravert is indicated by the SN

index. For introverts, the JP scale identifies the

auxiliary function, since the theory postulates that

introverts use the auxiliary function to relate to the

environment. By inference, the dominant function for an

introvert is the preferred function that wich is not

indicated by the JP scale to be extraverted.

Of the two components of the preference score, Myers

(1962) considers the letter indicating the direction of

preference to be more important than the numeric component.

She describes the numeric component as the "strength of the

preference." She

quantitative scores

that the respondent

31

cautions that interpretation of the

should be limited to the probability

has correctly reported his or her

preferences and the relative importance of these preferences

to him or her. She and numerous other researchers have

investigated correlates of the MBTI scales with other scales

of ability, aptitude, interests, and traits of personality

(e.g., Myers, 1962; stricker and Ross, 1964; McCaulley,

1978; Myers & McCaulley, 1985) reporting significant, but

generally moderate, associations in support of the

theoretical concurrent validity of the scales beyond their

use as categorical indices of Jungian types.

The interpretation of the quantitative score as a

measure of the "strength of the preference" is perhaps

confusing

taken to

in some instances where the term "strength" is

imply superiority of an ability to use the

function, or intensity of the preference for the function.

The procedures used in the item selections and the item

formats do not necessarily involve strength in either of

those connotations.

The preference scores are derived by summing the

responses in the direction of the preferred function, minus

the sum in the opposite direction, and applying a correction

when the scores are tied. This derivation seems more likely

to assess the "consistency" of the preference rather than

its "strength" or intensity. The difference here may be

32

largely semantic, particularly in light of Myers's cautions;

yet where the scale scores are used in correlational

studies, the term "consistency of preference" may lead to

less confusion. In this investigation, then, the magnitude

of the preference scores means the consistency with which

the respondents have indicated their use of the preferred

function among the opportunities provided by the

questionnaire to report them.

In selection of the items tested for inclusion in the

scales Briggs and Myers sought to avoid those that would

evoke extreme responses, favoring items that could reflect

the everyday manifestations of the type preference:

" . the questions can be trivial and often are" (Myers,

1961). The items sample a variety of situations in which

the response may be given easily from the respondent's

experience and with as little interference as possible from

the effects of social desirability or the requirement to

endorse extreme positions. Myers made no systematic attempt

to include items that would survey the full breadth of the

theoretical constructs (Myers, 1961); thus the items

retained through the selection process may reflect well some

aspects of the theoretical constructs and may not reflect

others. The evidence an item response predicts to the

expected direction sufficiently justified its retention.

The i terns are presented in two formats. Phrase

questions provide stems and alternative completions of the

33

stem. Word-pair items offer a choice between the attrac­

tiveness of the meaning of two words. In both formats the

respondent may choose only one of the responses offered, or

may omit the item.

Myers (1962) explains that the selection and testing of

the items and the construction of the scales and their

and 1956-1958) prior to the first publication of Form F in

1962. The aspects of the MBTI development history relevant

to the current investigation concern the adoption of item

weights and the testing of the item weights on samples of

differing ages.

History of the MBTI Item Weights

Briggs and Myers first performed item analyses of Forms

A and B of the MBTI on a sample of 114 male and 110 female

adul ts. The sample consisted mostly of college graduates.

Myers intentionally dropped individuals undergoing

psychological counseling and job applicants, in order to

exclude the biasing influence those circumstances might

introduce. The adult sample was selected it was presumed to

have greater differentiation of the type functions. Items

were retained if the response given by 60% of individuals

whose type (as determined by the total scale score) was in

agreement with the direction of the response. A second

analysis was later performed using the same criteria on four

samples: 248 adult males, 268 adult females, 224 college

34

males, and 70 college females. Only items reaching the

acceptance criteria for all of the four groups independently

were retained in Form C. The development of Form C

completed the first period of item selection and item

analysis.

In creating scoring procedures for Form C Myers

introduced separate weighting of each item response was

introduced in an effort to control for social desirability,

omissions, and unequal attractiveness of the response for

all the types sharing the preference. (At one point Myers

described detailed scoring procedures for each type but

never implemented them). The weights of the item responses

were then determined by a prediction ratio, designated PR,

and calculated as in the following formulae:

35

PRj = %J[J] / (%J[J] + %P[J]) and

PRp = %P[P] / (%P[P] + %J[P])

where

PRj

PRp

%J[J]

%J[P]

%P[P]

%P[J]

is the predictive value of the J response,

is the predictive value of the P response,

is the percent of judging types giving the

judging response,

is the percent of judging types giving the

perceptive response,

is the percent of perceptive types giving

the perceptive response, and

is the percent of perceptive types giving

the judging response.

In 1956 Myers undertooka new series of item analyses on

the Form C items, to which she had appended about 130

experimental items; the new questionnaire was known as Form

O. She undertook item weight refinement and item selection

with 385 graduate students to create Form 01, then with an

unreported number of undergraduate students to create Form

02. Finally, 2,573 high school students were sampled to

provide, as nearly as possible, 200 MBTI records for each

type, for each gender. The prediction ratio values

determined from these item analyses were used to set the

item weights for Form F: responses obtaining values from .63

to .729 scored one point; responses with values of .73 and

above scored two points. Responses falling below .63 or

36

responses having values in the second term of the

denominator greater than .5 scored zero points. The items

having one zero point response were retained following the

rationale that the non-zero alternative (assuming the

prediction ratio obtained a value greater than .63)

reflected an aspect of the type preference less popularly

endorsed and the endorsement of that response provided good

evidence of the type preference.

In shifting item weights from Form C to Form DO,

Briggs and Myers compared the scoring of college students to

that of the adults to ensure that item weight revisions did

not result in changes of type indication for the adult

criterion group. The item weights for Form DO were found to

be valid both for the adults and the 18- to 20-year-old

students. Similar checks of adult populations were not

reported for revisions from Form DO to Form Dl, Form Dl to

Form D2, or for Form D2 to Form F. In each of these

revisions, item analyses for progressively younger popula­

tions provided item weights which were adopted as the

standard for respondents of all ages.

In 1977, preparing for publication of Form G (which

contains simpler phrasing of some items), Myers conducted a

new item analysis on samples of students ranging from the

fourth to the twelfth grades and a sample of students just

entering college (Myers, 1977). These analyses showed that

the items , with the exception of a few, functioned among

37

students as young as the end of the fourth grade. That is,

they produced comparable i tem-to-scale correlations among

younger and older students. A second goal of the analyses

was to examine the prediction ratios to see if cultural

changes since the establishment of the item-weights 20 years

before had altered the popularity of some responses unduly.

The popularity of feeling responses was found to have

increased dramatically, and new item weights for the TF

scale were adjusted accordingly. Item weights established

in 1956 with the 11th and 12th grade students were retained

unchanged for the EI, SN, and JP scales. The new item

weights for the TF scale, derived from data obtained from

the students at the end of the fourth grade up to students

entering college, became standard for both Form G and

Form F. Myers (1977) reported that the new weights result in

reporting 5-15% more feeling types. The need to make these

item weight adjustments points to the difficulties of

calibrating the dividing points for the MBTI, since cultural

changes are likely to be reflected in the responses of

different generations.

The process of item selection and item weighting of

the MBTI began with small adult samples. Those items whose

responses correctly classified a type preference in the

criterion group were retained. In subsequent item selection

and weighting, the use of larger groups was necessary, and

for the type classification of these samples the criterion

38

was the scoring procedure of the previous version. The

weights currently being used were based mainly on the data

from matriculating college freshmen and high school and

middle school students. Also, regression analyses reported

by Myers (1962) in support of the correct placement of the

division points between the types were also based on high

school student data. These students were at the age that

Jacobi (1973) identified as having stable development of the

dominant function. Their reports of their preferences for

the dominant function, then, should have been reliable.

However, the students reports of their preferences for the

auxiliary function are likely to be less stable. No

investigators have reported analyses in which they made

this distinction--however, as noted above, Myers did explore

differential scoring procedures adjusted for the most

reliably reported preferences (Myers, 1961). Considering

the practical problems differential scoring procedures would

raise in manual scoring, that she did not implement them or

report them in detail is not surprising .

Current item weights may not be appropriate for

adults, especially for older adults. Even in the absence of

analyses indicating significant trends relating MBTI scale

scores to age, the consistency of the prediction ratios of

the i terns should be checked in older populations.

39

Review of the Research Literature: Psychological Type and Age

A search of the published MBTI literature cited through

May, 1984, revealed only one study that provided direct

evidence associating MBTI preference scores with the age of

respondents. In a doctoral dissertation, Bloch (1978)

reports tests of hypotheses advanced by Lundberg (1975) in

which Lundberg asserts that Jungian typology represents a

developmental sequence comparable to a stage theory of

psychological maturation. Following Lundberg, Bloch argues

that introversion, with its focus on abstraction reflects

psychological integration superior to that of extraversion,

which "focuses attention on external, concrete objects"

(Bloch, 1978, p. 6). Similar arguments are offered for the

superiority of intuition over sensing, thinking over

feeling, and judgment over perception. This arrangement

resul ts in a hypothetical progression for extraverts from

the lowest to the highest developed types as follows: ESFP,

ESTP, ESFJ, ESTJ, ENFP, ENTP, ENFJ, ENTJ. A parallel

progression for introverts is suggested, i.e., ISFP, ISTP,

ISFJ, ISTJ, INFP, INTP, INFJ, INTJ. In following strictly

the relative superiorities of the dichotomous pairs, the

developmental sequence should alternate between extraverted

and introverted types. Lundberg acknowledges that such

alternation was not only unlikely, but could not be

reconciled with theory. Thus, Lundberg hypothesizes

40

parallel sequences with individuals shifting from the

extraverted sequence to the introverted sequence at a point

where such a shift met less resistance than the shift to the

next highest extraverted type. No explicit point could be

indicated where this shift might occur. In formulating

these developmental sequences Lundberg has apparently relied

on superficial or very limited definitions of the attitudes

and functions and has largely ignored the dynamic inter­

actions described by Jung. For example, Jung says expli­

citly that the inferior may never be fully assimilated in

consciousness: "It would be a hopeless task--which never­

theless has often been undertaken and as often has

foundered--to transform an inferior function into a superior

one" (Jung, 1971, p. 86). For several of the hypothesized

type shifts, e.g., ESFJ to ESTJ, the dominant feeling

function would be replaced by a dominant thinking function.

The sequence further violates the Jungian typology in its

postulation of the relative value of the attitudes and

functions.

Psychology:

directly to

Jung responds to a questioner in Analytical

Its Theory and Practice as though speaking

Lundberg: "I hope I did not give you the

impression that I was giving a preference to any of the

functions," and, "We have absolutely no criterion by which

we can say that this or that function in itself is best"

(1968, p. 61).

41

The irreconcilable differences between (a) Lundberg's

developmental hypotheses regarding MBTI types and (b) the

Jungian type theory on which the MBTI is based severely

limit the utility of Bloch's study as the basis for a

developmental model. The study is further distanced from

the theory in the methods of analysis employed. Bloch

tested his hypotheses that the EI, SN, TF, and JP continuous

scores would have different means for each age group

studied, with the means differing in the predicted

directions. Parallel tests were made of the means of the

separate preference scores for each age group. In addition,

separate regression analyses were performed for each of the

four continuous scores and each of the eight preference

scores. However, Jung (1971) emphasized the interactive

dynamic of the attitude and function types: "strictly

speaking, there are no introverts and extraverts pure and

simple, but only introverted and extraverted function types,

such as thinking types, sensation types, etc." (p. 523). If

we are to test Jung' s typology, it is important to retain

the interactive nature of the preferences as far as

possible. Thus, to test the association of the Jungian

types with the age of the respondents, Bloch minimally would

have had to perform separate analyses for introverted and

extraverted function types for each of the preference

scores, e.g., separate analyses for introverted sensing

types and extraverted sensing types. The result of the

42

analysis design used by Bloch has been to confound the

frequencies of the preference types and the consistency of

preferences reported.

However, we may disregard the problems with the

correspondence of Bloch's (1978) theoretical formulation to

Jungian theory and examine the results of Bloch's analyses

simply for empirical evidence of an association between the

respondents's ages and the MBTI scale scores. Bloch

carefully selected his sample to avoid demographic bias in

his sample of 266 males and 292 females; he found no

significant differences for variables of gender, race, and

occupation between his sample and norms. The source of the

norms was not identified. He made no attempt to control or

select for the distribution of the type preferences within

the age groups examined. He reported statistically

significant tests for mean preference scores for sensing

and judging, both results having a probability less than

.001. Further, on additional regressions of the preference

scores he found trends showing that judging, sensing, and

intuition preference scores increase with age. Bloch tested

for the homogeneity of variance for the continuous scores,

and rejected the null hypothesis for the SN scale. He

suggests that his findings may have been influenced by the

lack of homogeneous variance. The study is important for

being the first in which investigators report systematic

association of the MBTI scale scores with the ages of the

43

respondents. The results generally run counter to Bloch's

developmental hypotheses but indicate a need for further

investigation. The finding that preferences for both

sensing and intuition increase with age is consistent with

Jungian theory and the results this writer found in a pilot

analysis of MBTI data. Failure to find similar associations

with the other five preferences cannot stand as evidence

that the associations do not exist.

The Isabel Myers Memorial Library housed at the Center

for Applications of Psychological Type contains numerous

research reports not available in the published research.

Among these reports is a paper by Schaeffer (1974), in which

she compares the frequencies of the type preferences in two

groups of "adult extension students". On chi-square test

extraverts were more frequent among the 30 students over age

35 than among the 170 younger students. She did not report

analyses of preference scores. Because of the small sample

size and the restricted population sampled Schaeffer's

findings would probably not replicate.

Because the two studies just described were the only

data found in which the investigator attempted to examine

the relationship of the MBTI scores to age, I extended the

literature search to include studies of the Jungian Type

Survey (JTS) (Gray & Wheelwright, 1946). The JTS, 16th rev.

(Wheelwright, et aI, 1964), is a 75-item self-report

questionnaire developed independently of the MBTI by two

44

Jungian analysts, Horace Gray and Joseph Wheelwright. They

obtain scores on three scales that are constructed to

reflect extraversion-introversion, sensing-intuition, and

thinking-feeling. In scoring they do not use item weights

and they make type classifications solely on the percentage

of responses in the direction of the preference. Gray and

and Wheelwright determine which function is the dominant

function by its being the function preference having the

higher percentage of responses endorsed. The JTS does not

have a scale corresponding to the JP scale of the MBTI;

thus, correlations with the JP scale are not available.

stricker and Ross (reported in Myers, 1962) obtained high

correlations for the three comparable scales among 47 male

college students. Myers (1962) concludes that after

adjustment for attenuation the correlated scales have their

entire true variance in common. Thus, reports of

correlations of the JTS with age are considered relevant to

the present investigation.

Gray (1946) reports a study of 500 male and 500 female

subjects who responded to the tenth revision of the Jungian

Type Survey. He found that when the subjects were divided

into five-year age groups, regression analyses showed that

the percentages of subjects classified as introverts,

sensing types, and thinking types increased with age.

Regressions of the mean scale scores for each age group also

indicated significant trends for increased introversion and

45

sensing; the trend for increased thinking fell just short of

the .05 probability level. He reports that he considered

each scale separately and did not distinguish between scores

of male and female respondents.

In a later report Gray (1948) addresses the question of

gender differences among the type preferences for the same

sample and reported males to be more often introverted,

sensing, and thinking (at probability levels of .09, .03,

and .01, respectively). Further analyses indicated that the

percentage of subjects with a perception function (sensing

or intuition) dominant over a jUdging function (thinking or

feeling) did not differ significantly between the sexes

overall; nevertheless, with the genders pooled he found that

the preference for a perception function became more

frequent with increased age. Despi te having found these

differences separately relating the dominance of type

functions to gender and to age, he does not report any

analyses to determine whether the relationships associated

with gender may not be constant across the age groups. His

data may suggest that gender and age do interact in their

effect upon the respondents I reports of their type

preferences. Males separately and older respondents (males

and females combined) both reported a greater preference for

introversion, sensing, and thinking. Among the same

subj ects the percentage of respondents (males and females

combined) preferring a perceptive function (sensing and

46

intuition combined) increased with age over the percentage

preferring a judging function (thinking and feeling

combined) . A set of regression lines that would satisfy

each of these conditions simultaneously, and would not

involve significant interactions of age and gender, is

difficult to imagine.

Thus, two objections to the adequacy of Gray's

analyses are raised here. First, in considering each of the

scales separately, Gray ignored Jung's theoretical position

that "strictly speaking, there are no introverts and

extraverts pure and simple, but only introverted and

extraverted function types, such as thinking types,

sensation types, etc." (Jung, 1971, p. 523). Further, the

separate analyses of the main effects of gender have left

unexamined a possible interaction of those factors.

Driver (1974) includes the Jungian Type Survey in a

study of personality variables related to age, types of

retirement residence, and length of stay in the residence.

The 185 subjects in her sample were 55 years or older (mean

age = 76.5 years) . Analyses of variance indicated

significant increases in the preference for introversion for

older groups, and showed that the introversion scale was

positively correlated with age, while the extraverted

preference had a negative correlation. Results with respect

to sensing, intuition, thinking, and feeling were not

47

significant. She does not report the type distribution of

her sample or gender-by-age analyses of JTS score means.

Before conducting the literature search summarized

above, I conducted exploratory analyses of 52,848 MBTI,

Form F, records contained in the MBTI data bank maintained

by the staff of the Center for Applications of Psychological

Type (CAPT). These records included all Form F cases

submitted to the CAPT MBTI computer scoring service from the

adoption of the current item weights for computer scoring in

1978 through 1982, with the exception of records meeting two

criteria. Records were excluded if 35 or more items had

been omitted (about 1% of the sample). In earlier analyses

of the CAPT data bank investigators found that the split­

half reliability coefficients for the MBTI scales dropped

markedly for subjects having more than 35 omissions (Kainz &

McCaulley, 1980). I also excluded records if the respond­

ents's ages were unavailable or were below 15 years.

For the purposes of these analyses the data were

divided into eight age groups: 15-17, 18-20, 21-24, 25-29,

30-39, 40-49, 50-59, and 60 or over. These age groups were

selected in part according to the age distribution found in

the data bank. The frequencies of respondents in the

younger groups were greater, so the younger groups had a

narrower age range for two reasons. First,was the desire to

separate young people still in high school from college

students and from young adults who were beginning their

48

careers. The second reason was an attempt to make the

frequencies of respondents in each group less

disproportionate.

I performed separate one-way analyses of variance for

each gender for each of the eight MBTI preference scores. I

found significant r ratios were found for each of the

preference scores (see Appendix A, Table 42). However, the

resul ts were not parallel for males and females. All r ratios for females were significant beyond the probability

level of .0001. For male extraverts and male feeling types

the probability of the r ratio fell just short of the .05

level. Post hoc tests to identify the significant contrasts

were not performed.

Visual inspection of the means (see Appendix A, Figures

12 through 15) suggests strong trends for increases in MBTI

preference scores. The preference scores for sensing,

intuition, thinking, and judging tend to increase with age

for both genders at least to age 50. The perceptive

attitude and the extraverted attitude appear to decline

among females, while only the perceptive attitude declines

for males. No increase of reported feeling preference

emerges for males, and the distribution of female feeling

scores appears to be bi-modal with a dip at the 25-29 group.

Also interesting is a possible change in the direction of

the trends at either the 40-49 or the 50-59 group,

especially noticeable for intuition and thinking. These

49

findings suggest that systematic trends are to be found

between respondent's age and MBTI preference scores that are

moderated by the respondent's gender.

These exploratory analyses have several technical

inadequacies that limit their interpretation. The

frequencies of respondents varied widely among the age

groups, with many more data in the younger groups; the

distribution of types varied from group to group, thus

allowing interactions among the type preferences to

influence the preference score studied, and the position of

the function as the dominant or auxiliary was not taken into

account.

Summary of the Problem

The Myers-Briggs Type Indicator was constructed

explicitly to assess preferences for the psychological

atti tude-functions described in Jung' s theory of psycho­

logical types. Previous theoretical discussions have not

suggested that MBTI preference scores would be

systematically related to the respondent's age in a manner

that is congruent with Jungian theory. Exploratory analyses

I performed on a large sample of MBTI records are suggestive

of systematic associations. Three of the four function

preferences (sensing, intuition, and thinking) appeared to

increase with age. The distribution of the preference

scores of the fourth function (feeling) differed by gender:

males did have a consistent increase in scores that did not

50

reach statistical significance; females seemed to produce a

bi-modal distribution having a valley at the 25-29 year

group. The attitude preferences suggested a systematic

increase in introversion and the judging attitude, with

corresponding decreases for extraversion and the perceptive

attitude.

In this review of the literature I found one writer

(Bloch, 1978) reporting increasing preferences for sensing

and judging among older respondents. An unpublished paper

by Schaeffer (1974) contained the report that extraverts

occurred more frequently in a small and biased sample of

students over 35 years old.

In a review of the literature regarding another Jungian

type indicator, the Jungian Type Survey, two reports by Gray

(1947b, 1948) contained reports of age and gender trends for

the JTS scores that partly paralleled the trends suggested

by my exploratory analyses. A study of elderly respondents

to the JTS (Driver, 1974) also associated introversion with

increased age. In all the analyses reviewed investigators

had problems with biased sampling and in every instance had

confounded the preferences of dominant and auxiliary

functions. The presence of significant findings across

several analyses suggests, nevertheless, that real trends

may exist that relate the Jungian type preferences to the

age and gender of the respondent.

51

Review of the Jungian type theory indicates that Jung

believed the type functions to be dynamically interrelated

in the development of ego-consciousness. The period from

childhood to adolescence is characterized by the

consolidation of one function and its associated attitude.

In adolescence or early adulthood there emerges in the ego­

complex an auxiliary, or balancing, function that differs

from the dominant by being a function (of either perception

or judgment) not served by the dominant, and by taking the

attitude opposing (and complementing) the dominant. The

attitude of the auxiliary is the attitude of the unconscious

represented in the ego. Through the auxiliary function,

consciousness becomes capable of regression in the service

of the ego, allowing previously unconscious conflict to be

consciously represented and managed within the ego-complex.

This development is the first step in the process of

individuation by which the "self" replaces the ego as the

center of conscious activity. At mid-life, the

individuating shift of consciousness from ego to self

reduces the relative importance of the dominant function and

increases the importance of the auxiliary function as the

mediator between the ego-complex and the broadening scope of

consciousness.

This sequence of development of the functions leads to

hypotheses relating the consistency of the MBTI preference

scores with these theoretically different periods of

52

psychological development. with respect to the dominant and

auxiliary functions, three periods are relevant: birth to

adolescence, adolescence to mid-life, and mid-life to death.

During the first period, the preference for the dominant

attitude-function combination increases markedly, with the

greatest discrepancy between the dominant and auxiliary

preferences occurring during late adolescence or early

adulthood. During the second period, the dominant attitude­

function combination continues to develop, but not so

rapidly as the auxiliary attitude-function combination

develops. At mid-life and beyond, the dominant attitude­

function does not further increase and may decline; however,

the auxiliary attitude-function increases, though at a

slower rate than earlier. The third and fourth functions

(the functions not reported as preferred on the MBTI) are

not expected to show any change in the consistency of their

preference as a function of consciousness.

The pattern just described models the optimal psycho-

logical development theorized by Jung. However, Jung and

his followers reported clinical observations that a number

of individuals fail to develop a stable auxiliary attitude­

function and reach mid-life unprepared for the task of

further psychological development which is described by Jung

to be marked by the shift from the ego to the self as the

focus of consciousness. For a number of others, the demands

of individuation beyond mid-life exceed the limits of their

53

capability. For those who respond to these demands by

broadening of consciousness, turning back from the

consciousness is decreased and narrowed as physical and

psychical energy declines. Reminiscences of the achieve­

ments and pleasures of the past crowd out the future and

much of the present. Therefore, we are warned by the

clinicians that we must expect, in some cases, a retreat, in

a desperate ego-centricity, to the narrower limits of the

dominant attitude-function (Jung, 1968, 1971; Jacobi. 1973).

In considering the connection between MBTI preference

scores and age, an alternative potential cause of systematic

association appeared in the history of the item selection

and development of the item weights for the scales. The

MBTI has undergone a series of revisions in which the

respondents in the normative samples used for selecting

valid items and item weights have been progressively

younger. The current weights used for respondents of all

ages were tested on a sample which was largely between the

ages of 16 and 18. The weights for the EI, SN, and JP

scales are three revisions away from those tested on adults,

and the TF scale is four revisions removed. In item

selection those items that discriminated the type

preferences of older individuals may have been discarded in

favor of items performing better among the samples in late

adolescence. For later research Myers retained in the

questionnaire most of the items that were deleted from the

54

scoring procedure for type preferences by the revision from

Form D2 to Form F. Some of these may prove to be valuable

with older respondents. It is also possible that the

weighting of the items that were retained may be more

effective among younger respondents than among older

respondents. If evidence of the latter were found,

alternative weights that are more appropriate for adult

respondents could be derived for the current questionnaire .

If evidence of the former situation were found, valid

weights may be determined for the items that are not now

scored.

The problem I addressed in this research was to

determine if the consistency of preferences for Jungian type

functions changes systematically with the age of the

respondent. To this end,the Jungian typology is operation­

ally assessed through responses to the MBTI. The consist­

ency of the type preferences is assessed by the magnitude of

the MBTI preference scores. Statistical associations found

between the MBTI preference scores and age may be an

artifact of the item selection and item weights of the MBTI.

Unfortunately, this possibility necessarily confounds

analyses of theoretically hypothesized relationships. In

this research separating these effects or controling

statistically for this potential confounding is not

possible. However, separate analyses may be performed that

55

test the extent to which the analyses of preference scores

and age are influenced by the item weights.

Ideally, this problem would be addressed by analyses

of longitudinal data. However, suitable data do not exist

and cannot reasonably be obtained. The best alternative is

a large sample of cross-sectional data with controlled

sampling for gender, age, and type preferences. A suffi­

ciently large sample has been generated through the MBTI

computer scoring service of the Center for Applications of

Psychological Type. Those data now include more than 75,000

MBTI Form F records obtained between 1978 and 1984.

The first step in determining trends associating the

MBTI preference scores with the respondent's age was

addressed by randomly selecting equal numbers of records for

each type and gender at each age group from the CAPT MBTI

data bank. The number of records selected for each

type/gender/age was determined by the number of

available to the smallest group to be included

particular analysis. The initial pool of records

records

in the

in the

Form F data bank was greater than 82,000. Excluding records

with excessive item omissions, no recorded age, or age less

than 15, left more than 75,000 potential observations. The

actual sample size analyzed was considerably smaller due to

the marked decrease in the frequency of records from younger

to older respondents. I examined these records of the

equal-type research sample by analyses of variance and

56

linear regression analyses for differences in mean

preference scores and for linear association. I identified

subsets of the data sample by dominant or auxiliary

functions, by the eight age groups used in the piolot

analyses, and by age period (15 to 39 years, or 40 years and

over), and by gender. The analyses of main effects and

interaction effects of these factors were used to test

research predictions derived from Jungian theory relating

the differentiation of the dominant and auxiliary functions

to the ages of the respondents.

The same MBTI records used in the regression analyses

were used to examine age differences in MBTI item weights.

Following the procedures described by Myers (1962), the

prediction ratio (PR) of each response was calculated for

each age group catagorized by sex. Correctly weighted item

responses were expected result in calculated prediction

ratios for each group that lie in the range used by Myers

for the assignment of the weight. Responses showing

deviation from the expected range were noted for further

study. The extensive evaluation which would be required to

establish alternative item weights was beyond the scope of

the present study. The purpose of this item analysis was to

determine if there are items which may be candidates for

such an evaluation. The presence of item responses that

have predictive value that varies with the ages of the

respondents was expected to influence the findings of

57

analyses of variance and analyses of trends of association

relating respondents's ages to MBTI preference scores.

Research Predictions

In this section the model of development of the Jungian

personality types discussed above will be related to

research predictions associating trends in the MBTI

preference scores with the respondent's age. In the Jungian

model three general periods of development may be

identified: (a) the development of a dominant attitude­

function, (b) the development of an auxiliary attitude­

function, and (c) individuation, or the development of self­

consciousness from ego-consciousness. The first period

occurs from birth to approximately mid-adolescence; the

second period, from mid-adolescence to mid-life; the third

period, from mid-life to death. This research is limited to

the latter two periods because they are the age ranges for

which the MBTI may be reliably used, that is, 15-39, and

from 40 on. The division between the two periods is

arbitrary, lying within the 35-to-45-year range identified

by Jung as the beginning of mid-life.

During the second period (age 15-39) it is expected

that preference scores reflecting the consistency of the

preference for the dominant attitude-function should display

a low yet steady increase with aging; preference scores

reflecting the consistency of the preference for the

auxiliary attitude-function should show a greater tendency

58

to increase with age, relative to the increase of the

dominant attitude-function. During the third period the

preference scores reflecting the preference for the dominant

attitude-function should not increase, and may decline;

however, preference scores reflecting the consistency of the

preference for the auxiliary attitude-function should

continue to increase with age.

The possible combinations of dominant and auxiliary

attitude-functions describe 16 personality types. Any

single type shares the dominant attitude-function with one

other type, which always has a different auxiliary attitude­

function. This relationship is more easily presented in

specific examples. For instance, the MBTI type formula ENFP

indicates preferences for extraversion (E), intuition (N),

feeling (F), and perception (P). The preference to orient a

function of perception by extraversion is indicated by the

fourth letter, P. Being an extravert, the individual uses

the dominant function to relate to the external world.

Therefore, the dominant attitude-function of the ENFP is

extraverted intuition. By extension, the auxiliary

attitude-function of the ENFP type is introverted feeling.

The ENFP type may also be described as an extraverted

intuitive type with introverted feeling. The ENTP type is

also an extraverted intuitive type, though for the ENTP the

auxiliary attitude-function is introverted thinking. The

preference for the dominant is theoretically independent of

59

the preference for the auxiliary, except that the auxiliary

may not be directly opposed to the dominant. For analyses

pertinent to trends of the preference for extraverted

intuition as a dominant attitude-function, ENFPs and ENTPs

may be pooled so long as both types are equally represented.

To pool these two types in analyses related to the auxiliary

atti tude-function would not be appropriate. However, the

ENFP type has the same auxiliary attitude-function as the

ESFP type, i.e., introverted feeling. In analyses of

introverted feeling as an auxiliary attitude-function the

statistician may, therefore, pool these two types. Table 1

displays the dominant and auxiliary attitude-functions

corresponding to each of the 16 MBTI type formulae. Table 2

displays the predicted direction and intensity of change of

the MBTI preference scores for each of the 16 types as they

age.

In the example of the ENFP type, the preference scores

reflecting the consistency of the preference for extraverted

intuition as the dominant attitude-function include those

for extraversion, intuition, and perception. The model then

predicts that these preferences should increase with the age

of the respondent during the second development period. For

the ENFP type, the preference scores reflecting the

consistency of the preference for introverted feeling as the

auxiliary attitude-function include those for introversion

and feeling. The model predicts a greater increase in these

60

preferences during the second development period than the

increase predicted for the dominant attitude-function.

Thus, for the ENFPs, the preference scores for intuition and

feeling should increase with age; feeling, the auxiliary

function, increases more rapidly during the second period

than does intui tion. Note that the preferences for

extraversion (associated with dominant intuition) and

introversion (associated with auxiliary feeling) are both

predicted to increase. However, during this period the

development of the auxiliary is more rapid. Therefore,

introversion ,associated with the auxiliary, is predicted to

increase more rapidly than extraversion, which is associated

with the dominant. Due to the method used to compute MBTI

scores, an increase in the reported preference for one

attitude (or function) necessarily produces a decrease in

the reported preference for its opposite. Thus, in

instance the result should be a net decrease in

extraversion preference score.

this

the

The preference score for perception as the attitude

used for orienting consciousness to the external (or extra­

verted) world is associated, for the ENFP type, with the

continuing development of intuition. Nothing in the theory

predicts development of an extraverted judging function in

the ENFP; therefore the result is an increase in the

perception preference score with age.

61

During the third period (age 40 and beyond), the model

predicts that the preference for the dominant attitude­

function should not increase, and the preference for the

auxiliary attitude-function should increase. Thus, for the

ENFP, the intuition preference score should remain constant

with age, and the feeling preference score should continue

to increase. The preference for extraversion, as with

intuition, stays constant; however, introversion, as with

feeling, increases. Again the result should be a decrease

in the extraversion preference score. The perception

preference score should stay constant in parallel with

intuition. These predictions are summarized in Table 2.

I have arranged the order of the types in Table 2 such

that four groups (each containing four types) share

identical patterns of predicted changes. Jung identified

these groups as extraverted rational types, extraverted

irrational types,

verted irrational

introverted rational

types, respectively.

types, and intro­

Jung used these

terms to denote the fact that the four types within each of

the groups share a common aspect of the dominant attitude­

function. The two groups (one extraverted and one

introverted) termed rational by Jung both have a judging

function dominant. The two groups termed irrational by Jung

both have a perceptive function dominant. The four types

associated with each group are listed in Table 3.

62

The patterns for the extraverted types and the

introverted types are symmetrical except for the JP scale.

The lack of symmetry arises because the JP scale reflects

development of the function oriented by extraversion.

Thus, predictions for change in

judgment or perception preference

the consistency of the

scores must follow the

extraverted. For extraverts the dominant function is

extraverted, while for introverts the auxiliary function is

extraverted.

Table 2 shows that the same predictions are made at

each stage for all of the extraverts and for all of the

introverts with respect to changes in the EI preference

scores. Likewise, within each of the four rational­

irrational groups, identical predictions are made for

changes in the JP preference scores. Further, for the pairs

of types that share the same dominant or the same auxiliary

function, identical predictions are made for the shared

function. Types sharing the same dominant or auxiliary

attitude-function are listed in Table 4.

The research predictions shown in Table 2 for each

preference of each of the 16 types may be reorganized as in

Table 5 according to the type groups that have an identical

prediction for a preference score.

63

Table l. Dominant and Auxiliary Attitude-Function of MBTI Types

MBTI Type Dominant Auxiliary

ESTJ extraverted thinking introverted sensing

ENTJ extraverted thinking introverted intuition

ESFJ extraverted feeling introverted sensing

ENFJ extraverted feeling introverted intuition

ESTP extraverted sensing introverted thinking

ESFP extraverted sensing introverted feeling

ENTP extraverted intuition introverted thinking

ENFP extraverted intuition introverted feeling

ISTP introverted thinking extraverted sensing

INTP introverted thinking extraverted intuition

ISFP introverted feeling extraverted sensing

INFP introverted feeling extraverted intuition

ISTJ introverted sensing extraverted thinking

ISFJ introverted sensing extraverted feeling

INTJ introverted intuition extraverted thinking

INFJ introverted intuition extraverted feeling

64

Table 2. Predicted Changes of MBTl Preference Scores by Types

MBTl Second Development Period Third Development Period Type (age 15 to age 39) (age 40 and older)

ESTJ E(- S(++) T(+ J(+ E(- S(+ T(= J(=

ENTJ E(- N(++) T(+ J(+ E(- N(+ T(= J(=

ESFJ E(- S(++) F(+ J(+ E(- S(+ F(= J(=

ENFJ E(- N(++) F(+ J(+ E(- N(+ F(= J(=

ESTP E(- S(+ T(++) P(+ E(- S(= T(+ P(=

ESFP E(- S(+ F(++) P(+ E(- S(= F(+ P(=

ENTP E(- N(+ T(++) P(+ E(- N(= T(+ P(=

ENFP E(- N(+ F(++) P(+ E(- N(= F(+ P(=

lSTP l(- S(++) T(+ P(++) l(- S(+ T(= P(+

lNTP l(- N(++) T(+ P(++) l(- N(+ T(= P(+

lSFP l(- S(++) F(+ P(++) l(- S(+ F(= P(+

lNFP l(- N(++) F(+ P(++) l(- N(+ F(= P(+

lSTJ l(- S(+ T(++) J(++) l(- S(= T(+ J(+

lSFJ l(- S(+ F(++) J(++) l(- S(= F(+ J(+

lNTJ l(- N(+ T(++) J(++) l(- N(= T(+ J(+

lNFJ l(- N(+ F(++) J(++) l(- N(= F(+ J(+

NOTE: + increase with age ++ greater increase with age decrease with age = constant with age

65

Table 3. MBTI Types Grouped by Jung as Rational and Irrational

Group Types Included

Extraverted Rational ESTJ, ESFJ, ENTJ, ENFJ

Extraverted Irrational ESTP, ESFP, ENTP, ENFP

Introverted Rational ISTP, ISFP, INTP, INFP

Introverted Irrational ISTJ, ISFJ, INTJ, INFJ

Table 4. MBTI Types Sharing Dominant or Auxiliary Attitude-Functions

Attitude-Function Dominant in Auxiliary in Types: Types:

Extraverted Thinking ESTJ, ENTJ ISTJ, INTJ

Extraverted Feeling ESFJ, ENFJ ISFJ, INFJ

Extraverted Sensing ESTP, ESFP ISTP, ISFP

Extraverted Intuition ENTP, ENFP INTP, INFP

Introverted Thinking ISTP, INTP ESTP, ENTP

Introverted Feeling ISFP, INFP ESTP, ESFP

Introverted Sensing ISTJ, ISFJ ESTJ, ESFJ

Introverted Intuition INTJ, INFJ ENTJ, ENFJ

66

Table 5. Predicted Changes of MBTI Preference Scores, by Groups of Types

Type Group Predicted

Preference Score

Developmental Period

Second Third

Attitudes

Extraverts E

Introverts I

Extraverted Rational J + =

Extraverted Irrational P + =

Introverted Rational P ++ +

Introverted Irrational J ++ +

Dominant Functions

Extraverted Thinking T + =

Extraverted Feeling F + =

Extraverted Sensing S + =

Extraverted Intuition N + =

Introverted Thinking T + =

Introverted Feeling F + =

Introverted Sensing S + =

Introverted Intuition N + =

67

Table 5. continued.

Type Preference Developmental Period Group Score Predicted Second Third

Auxiliary Functions

Extraverted Thinking T ++ +

Extraverted Feeling F ++ +

Extraverted sensing S ++ +

Extraverted Intuition N ++ +

Introverted Thinking T ++ +

Introverted Feeling F ++ +

Introverted Sensing S ++ +

Introverted Intuition N ++ +

NOTE: + indicates increase with age ++ indicates stronger increase with age

indicates decrease with age = indicates constant with age

CHAPTER II METHODS

Introduction

This chapter is a description of the data sample

selection and the methods of data analysis used in this

study. The problem, as discussed in the previous chapter,

required a large sample of MBTI records which was equally

representative of each of the 16 MBTI types and of each

gender at each of the age levels selected for study. I

analyzed the data in two separate ways: (a) analyses of

variance and regression analyses on the preference scores by

age groups and by age respectively for each gender, and (b)

item analyses of the prediction ratios at each of the age

groups for each gender.

Selection of the Sample

The Center for Applications of Psychological Type

(CAPT) in Gainesville, Florida, maintains a databank con-

taining of all MBTI questionnaires submitted to CAPT for

computer scoring. Mary H. McCaulley and Isabel Myers

established the data bank to make available large samples of

MBTI records for psychometric and normative studies. Those

who submit MBTIs for computer scoring--individuals and

68

69

groups--are informed that the records are archived for

research requiring aggregated data and that the data they

submit will be so used. Data used in this study were drawn

from the CAPT MBTI records submitted between March, 1978,

and June, 1984.

In June, 1984, the CAPT data bank contained 82,498

MBTI Form F records. Respondents had omitted 35 or more

items on 637 (.08%) of the indicators. These records were

dropped because earlier studies done at CAPT (Kainz &

McCaulley, 1980) of split-half reliability co-efficients had

shown unacceptable reliability co-efficients for MBTI

scores from such records. The remaining 82,831 records were

sorted alphabetically by the respondents' names and

sequentially assigned a random number drawn from a file of

90,000 non-repeating random numbers; thus, each record had a

unique randomly generated identifier.

The MBTI respondent's age was computed by subtracting

the reported date of birth from the reported date of

testing or from the date the MBTI was scored if the date of

testing was not reported. After the deletion of 5,187 cases

for which no age could be computed, the resulting sample

contained 76,764 records.

These records were separated into nine age groups:

9-14, 15-17, 18-20, 21-24, 25-29, 30-39, 40-49, 50-59, and

60 or over. The youngest group (9-14 years) was dropped as

being below the age for which the MBTI is known to be

70

reliable. Members of the remaining eight age groups were

further separated by gender and type. This process produced

256 groups (8 ages x 2 genders x 16 types). Each of these

groups was placed in random order by sorting the records

into the sequential order of their randomly generated

identifiers. The frequencies of records varied widely

across types and gender wi thin the age groups and this

required further selection to balance the inequities.

The frequency of the smallest type-gender group at

each of the eight age levels determined the frequency for

all other type-gender groups at that age level. The

required number of records was selected sequentially from

each of the randomly ordered groups. This process achieved

the goal of randomly selecting, at each age level, the

largest possible sample size thatwas balanced for type and

gender. Table 6 displays the frequencies at each age level

of the final sample.

Potential Sources of Sampling Bias

The process for selecting the sample provided an equal

representation of types and gender at each age level, but

was limited by the records available in the CAPT data bank.

McCaulley (Myers & McCaulley, 1985) noted several factors

that limit the representativeness of the CAPT data bank with

respect to the general population. Most significant among

these is the relatively high education levels of the MBTI

respondents (31 report having completed at least some

71

college). The CAPT data bank and the data sample drawn from

it should not be considered to be representative of the

general population. For the purposes of this study such

representativeness is not a requirement. Regarding other

factors, however, the data bank sample is likely to be

representative of the populations with whom the MBTI is

currently being used.

Analyses of Variance and Regression Analyses

within the data sample the MBTI records weresub­

divided into 256 groups (2 genders x 8 age groups x 16

types), each containing the records for MBTI respondents

reporting the same type, gender, and age group. Sub-samples

consisting of groups of types were compiled. For each sub­

sample (hereafter referred to as sample) the MBTI preference

score treated as the dependent variable shared the same

attitude-function as the other types in the sample, e.g.,

dominant extraverted intuition. The MBTI records included in

each analysis were determined by the groupings of types

listed in Table 5. Separate analyses of variance and

regression analyses were performed for each gender. Thus,

analyses of the relationship of age to females preference

for extraversion included all records from female

extraverts. Likewise, analyses of the relationship of age

to males preference for extraverted thinking as an auxiliary

attitude-function included all males whose reported type

preferred extraverted thinking as an auxiliary function,

72

i. e., ISTJ and INTJ. For each analysis the type of MBTI

respondents included may be found in Tables 3 and 4.

Differences of mean preference scores were predicted

both (a) across the eight age groups and (b) across the two

theoretical stages of the development of type preferences

(age 15-39, and 39 and over). Also, a definite pattern of

trends of association between the preference scores and age

were predicted for the two developmental stages. Analyses

of variance were performed for both the eight age groups

(see Table 6) and the two theoretical stages of development

defined as the independent variable. Regression analyses

were performed across the full age range of the sample and

separately for the age ranges of the developmental stages.

For each analysis of variance, the null hypothesis

being tested was that no differences exist among the means

of the MBTI preference scores between the age groups, or

between the developmental stages. For each regression

analysis for trends of association, the null hypothesis

being tested was that no trend of association exists between

the MBTI preference scores and the ages of the respondents.

In Table 5 are listed the 22 samples of MBTI types on

which the analyses in this study were performed. For each

grouping of types, two analyses of variance and two

regression analyses are indicated for each gender. The

study, thus, includes 88 separate analyses for each gender

or 166 analyses in all. Because so many analyses were

73

performed, the null hypothesis was not rejected unless the

probability of the test statistic E was .005 or less.

Homogeneity of Variance in the Sample

The large variation in the frequencies of observations

available for each of the eight age groups raised consid­

erable concern that the analyses might be distorted by large

departures from the homogeneity of the variance of the

dependent measures. For each of the 22 type groupings the

homogeneity of variance of the relevant MBTI preference

score was tested across the age

stages used in the analyses.

groups and developmental

Analyses of the Item weights

Myers and McCaulley (1985) reported that the weights

assigned to the item responses scored on the MBTI are

determined according to a prediction ratio (PR) which is

calculated as follows:

PRj = %J[J] / (%J[J] + %P[J]) and

PRp = %P[P] / (%P[P] + %J[P])

where:

is the predictive value of

is the predictive value of

the J response,

the P response,

PRj

PRp

%J[J] is the percent of Judging types giving the

jUdging response,

%J[P] is the percent of Judging types giving the

perceptive response,

74

%P[P] is the percent of Perceptive types giving

the perceptive response, and

%P[J] is the percent of Perceptive types giving

the jUdging response.

Responses obtaining a calculated PR from .630 and .729

have a weight of one point. Responses obtaining a

calculated PR of .730 or above have a weight of two points.

Responses with calculated PRs below .630 or endorsed by more

than 50% of respondents reporting the opposite preference

(e.g., %P[J] or %J[P] greater than 50 percent) have a weight

of zero points. The zero point weight is assigned on the

assumption that responses that are generally too popular

provide little evidence of a type preference. The items

with zero point responses were retained because the

endorsement of the unpopular alternate response may provide

good evidence of a type preference.

The prediction ratios and popularity of each of the

item responses were calculated by the MBTI Item Analysis

Program (Grennade, 1985) for the total sample, for females,

for males, for all age groups combined, each age group

separately and each stage separately. Following Myers's

procedures in all analyses, records were dropped from the

calculation if the preference score on the relevant scale

was within one point of the mid-point of that scale.

omitted responses were not included in the calculations for

that item.

75

The assumption underlying analyses of the response

weights undertaken in this study is that Myers, in her

original method of response weighting, established an appro­

priate standard of predictive value for each response based

on the MBTI samples available to her, and established an

appropriate method (prediction ratios and popularity) to

assess a response's performance. Myers reported (1961) that

thousands of proposed item responses that failed to meet her

standard were discarded as inadequate. The central question

addressed in these analyses is whether the responses that

met the predictive criteria among the high school and

college students perform as well among older groups.

Responses that are not as predictive as expected by Myers's

criteria are considered over-weighted. Responses that

exceed the criteria are considered under-weighted.

Responses are considered over-weighted if the

calculated PR was less than the minimum PR specified by

Myers for the assigned weight of the response, or if the

popularity of the response exceeded 50%. Responses are con­

sidered under-weighted if the calculated PR was greater than

the maximum PR specified by Myers for the assigned weight,

and if the popularity of the response did not exceed 50%.

Responses are considered correctly weighted if the calcu­

lated PR was between the maximum and minimum PRs specified

by Myers for the assigned weight of the response, and if the

popularity of the response does not exceed 50%.

76

computer Software Used in Sample Selection and Data Analyses

All record manipulations and random selection of

records were performed by specially written computer

programs prepared at CAPT, or by utility programs contained

in the UNIX V. 7 operating system. All statistical calcu-

lations were performed by the SPSS/PC: Release 1.1 statis-

tical analysis software system (Norusis, 1984). Item

analysis calculations were performed by the MBTI Item

Analysis Program (Grennade, 1985).

77

Table 6. Sample frequencies per gender for age groups

Age Age N per N per N per Column Group Range Types Preference Age Group Percent

1. 15 - 17 84 672 1,344 12.4

2 . 18 - 20 214 1,712 3,424 31. 6

3 . 21 - 24 117 936 1,872 17.3

4. 25 - 29 88 704 1,408 13.0

5. 30 - 39 92 736 1,472 13.6

6. 40 - 49 56 448 896 8.3

7. 50 - 59 20 160 320 2.9

8. 60 and over 7 56 112 1.0

Total Females 678 5,424 10,848 100.1

Total Males 678 5,424 10,848 100.1

Total Sample 1,356 10,848 21,696 100.1

CHAPTER III RESULTS

Introduction

This chapter is a presentation of the results of the

analyses of variance and the regression analyses of the MBTI

preference scores by age of the respondents, and the results

of the item analyses. Chapter I included a review of

several studies in which investigators reported a 'statis-

tical association between MBTI preference scores and the

respondents's ages. The sampling procedures in those

studies and several investigators omission Jung's distinc-

tion between dominant and auxiliary attitude-functions were

cited as methodological deficiencies that limit the utility

of the results they reported.

I undertook the present study to correct those

deficiencies and to evaluate a set of research predictions

that attempt to relate changes in MBTI to age in a manner

consistent with Jung' s theory. A further purpose was to

examine the MBTI item weights to determine if the weights

are appropriate across the age range for which the MBTI is

used.

78

79

To achieve these purposes I selected a large sample

(N = 21,696) of MBTI records from the data bank of MBTI

records maintained at the Center for Applications of Psycho­

logical Type (CAPT). The selection procedures insured that

equal numbers of records were included for each of the 16

MBTI types for each gender within eight age groups. I

analyzed the differences among the mean MBTI preference

scores by age groups and by hypothesized stages of type

preference development, and trends of association between

the MBTI preference scores and age were analyzed by

regression analyses across the full age range and within the

age ranges of the developmental stages. I tested the

variances of the preference scores across the age groups and

stages used in the analyses for departures from homogeneity.

Finally, I replicated on the full age range the procedures

for item analyses employed by Myers (Myers & McCaulley,

1985) to determine the current item weights, and replicated,

also, for each age group and developmental stage separately.

Analyses of Variance and Regression Analyses

The results of the analyses of variance, tests for

departure from homogeneity of variance, and regression

analyses for trends of association will be presented in

separate sections for the attitude scales (EI and JP), the

function scales, the dominant attitude-functions (ET, EF,

ES, EN, IT, IF, IS, IN), and the auxiliary atti tude-

functions (ET, EF, ES,. etc.)

80

Due to the large number of analyses of variance and

regression analyses performed in this study the null hypo­

theses (of no difference among group means or no linear

trends of association) were rejected only if the test

statistics had a probability of .005 or less of occurring by

chance.

Extraversion and Introversion

The research predictions for the EI preference scores

(summarized in Table 5) were consistent for all types for

both developmental stages. Both preferences were predicted

to decrease with the age of the respondent, because during

the second developmental stage (age 15- 39), the preference

for the auxiliary attitude function is expected to have

greater increase than the preference for the dominant. Due

to the ipsative nature of the MBTI scales, an increase in

the auxiliary attitude also produces a decrease in the

dominant attitude. For all subjects preference for the

attitude of the dominant is always greater. Thus, an

increase in the consistency of the auxiliary is observed in

a decrease in the consistency of the attitude of the

dominant.

In Figure 1 the mean EI preference score was plotted

for extraverts and introverts of each gender by age groups.

By visual inspection, what variability is present appears to

be relatively small. Nevertheless, the stability of the

variance estimates from such a large sample permits

81

statistical analysis to detect even

effects. Table 7 is a summary of

small but reliable

the results of the

analyses of variance for the age groups. Each of those

tests indicates differences among the mean EI preference

scores for each of the samples. Bartlett-Box tests for the

homogeneity of variance indicate departure from homogeneity

among the variances for female extraverts. Analyses of

variance of the EI preference scores by developmental stage

(see Table 8) indicate that only female extraverts differ

significantly in the consistency of preference at different

stages. Variance estimates are homogeneous across the

developmental stages.

Tables 9, 10 and 11 summarize the results of regres-

sion analyses for trends of association with age. Across

the full age range, a small trend is indicated for female

extraverts to decrease their preference for extraversion and

for both male and female introverts to increase their

preference for

extraverts.

introversion. No trend occurs for male

During the second developmental stage (15-39 years),

trends occurring for the full age range are present with

slightly greater strength (see Table 10). It would appear

that the trends of the second stage largely account for the

trends found for all ages. For the third developmental

stage (age 40 and over) no statistically reliable trends

occurred (see Table 11).

82

Taken together the results show that the consistency

of preference for extraversion and introversion varies

reliably with the age of the respondent but only to a small

degree. The largest effect is among female extraverts

between the ages of 15 and 39, who report a decreasing

preference for extraversion. Analyses of variance show

reliable differences among the mean EI preference scores for

both males and females across eight age groups. However,

when the data are grouped by developmental stages, the

differences appear only among female extraverts. Among the

other samples the differences of the means lie more within

the stages than across them.

Al though younger extraverts did show a decrease in

reported extraversion over time, the overall trends are

more consistent with the interpretation that respondents

either do not change their reported preference for extra­

version or introversion or they increase their reported

preference for introversion as they grow older.

Judging and Perceptive Attitudes

The jUdging and perceptive attitudes are interpreted

as being the result of extraverting either a judging or a

perceptive function. According to Jung's theory of types,

the dominant function of extraverts will be extraverted and

the auxiliary function of introverts will be extraverted.

Predicted changes in the consistency of the preference for

the judging or perceptive attitude are the same as changes

83

predicted for the extraverted function, i. e., for extra­

verts, the dominant function, and for introverts, auxiliary

function. During the second developmental stage, the prefer­

ence for the dominant function is predicted to increase to a

small extent and the preference for the auxiliary function

is expected to increase to a greater degree. During the

third developmental period, the preference for the dominant

function is expected to remain constant and the preference

for the auxiliary function is expected to increase. There­

fore, during the second stage, the JP preference score of

extraverts is expected to increase to a small degree and

the JP preference score of introverts is expected to have a

greater increase. During the third stage the JP preference

scores of extraverts is expected to remain constant and the

JP preference score of introverts is expected to increase to

a small degree.

The analyses were performed separately for samples

having the same predicted changes and the same jUdging or

perceptive preference. Jung termed these groups: extra-

verted rational, extraverted irrational,

rational, and introverted irrational.

introverted

Figures 2 and 3 display the mean JP preference scores

across the eight age groups. By visual inspection it is

apparent that among both genders the samples reporting a

preference for the judging attitude (extraverted rational

and introverted irrational types) report a more consistent

84

preference for jUdging among the older age groups. Among

the samples reporting a preference for the perceptive

atti tude (extraverted irrational and introverted rational

types) the reported preference for perception apparently

decreases among the older groups for extraverted irrational

types of both genders, and for male introverted rational

types. Among the female introverted rational types the

preference for the perceptive attitude appears to decrease

slightly from the first through the fourth age group, to

increase about the same degree through the seventh age

group, and finally decline to the level of preference

reported by the male introverted rational types.

Analyses of variance by age groups, summarized in

Table 12, confirmed reliable differences among the mean JP

preference scores for the seven samples. The mean JP

preference

introverted

scores do not differ by age among the female

rational types. Analyses of variance by

developmental stages, summarized in Table 13, also indicate

differences that are statistically reliable. Five of the

eight samples show differences at the .005 probability level

or beyond. All but the introverted rational types and the

male extraverted irrational types differ in the preference

for the perceptive attitude between the second and third

developmental stages. The fourth sample reporting a prefer­

ence for the perceptive attitude (Le., male introverted

rational types) does not have homogeneity of variance of the

85

JP preference scores at the two developmental stages.

Therefore, the results for that sample may be distorted.

The analyses of trends across the full age range

summarized in Table 14 show that the differences detected by

the analyses of variance have a reliable linear association

for all samples except female introverted irrational types.

Each of the samples reporting preference for the judging

attitude have reliably positive associations between age and

the JP preference score. Three of the samples preferring the

perceptive attitude have reliable, though smaller, negative

associations between age and the JP preference score. The

trends for the second developmental stage, shown in Table

15, are similar to those for the full age range. However,

for each sample having reliable associations, the trend is

weaker. While some small trends continue into the third

stage, none are statistically reliable beyond a probability

of .005 (see Table 16).

The pattern of trends of association between the

consistency of preference for the judging or perceptive

atti tude do not follow the expectations presented above.

Negative trends occur among the extraverted irrational types

and male introverted rational types, for whom a positive

trend was predicted. The results obtained are more

consistent with another interpretation: that for those

samples reporting a preference for the judging attitude, the

preference is reported more consistently at older ages;

86

whereas, among those samples reporting a preference for the

perceptive attitude (except female introverted rational

types) the preference is reported less consistently at older

ages.

It is of interest to note the differing results

obtained for male and female introverted rational types.

Type theory provides no basis for predicting gender differ­

ences between these two samples. Much of the gender differ­

ence observed (Figure 4) is accounted for by the relatively

lower preference for the perceptive attitude reported by

female introverted rational types in the lowest age groups,

when compared to male introverted rational types and to all

extraverted irrational types.

Figures 3 and 4 (although

reliability) that females who

It may also be observed from

not tested for statistical

report a judging attitude

report a more consistent preference at the lowest age groups

than do males who prefer jUdging.

Dominant Sensing-Intuition

Parallel research predictions were made for all of the

attitude-functions. The second developmental stage (age

15- 39) is theoretically the period of greatest development

of the auxiliary attitude-function and, to a lesser extent,

of further development of the dominant attitude-function.

During the third developmental stage (age 40 and over), the

demands of individuation require a lessening of identifi­

cation of ego consciousness with the dominant attitude-

87

function and an increasing interest in previously undevel­

oped attitude-functions. Therefore, it was predicted that

the MBTI preference scores related to the dominant attitude­

function would show a small yet reliable increase during the

second developmental stage, and then would remain constant

during the third developmental stage.

Figures 4 and 5 display plots of the mean SN

preference scores by age groups for those samples having a

dominant perceptive attitude-function. By visual inspec­

tion, distinct patterns of association appear for each of

the attitude-function samples, patterns that are largely

shared by both genders. Among the extraverted samples there

is a general tendency for the preference to increase slowly

through the age groups of the second developmental stage

(groups I 5), with the exception of a drop in the

preference for extraverted intuition among the 18-20-year­

olds. Among extraverted respondents aged 60 and older,

females who prefer extraverted sensing show a continued

increase in their preference, but the preferences of the

other extraverted samples decrease. Among introverts, those

preferring sensing display a clear tendency to increase

their preference in the older age groups. The introverted

intuitives increase their preferences as well, but to a

lesser degree. Also of note are the consistent shifts in the

level of the reported preference by attitude-function

samples. Extraverts preferring intuition prefer it more

88

than extraverts preferring sensing prefer sensing. Among

introverts the relationship is opposite and even more

apparent.

means.

All of the sensing means exceed the intuition

The analyses of variance of the SN preference scores

by age groups, summarized in Table 17, indicate reliable

differences among the means for all samples except

extraverted sensing. However, Bartlett-Box tests indicate

that significant departures from homogeneity of variance

exist among the introverted sensing samples for both

genders. The differences remain in mean SN preference

scores for extraverted intuition among males and for intro­

verted sensing, when the data are grouped by developmental

stages (see Table 18). The differences among the means for

the introverted intuitive samples appear to lie within the

stages rather than between them.

The regression analyses for trends of association

between the SN preference score and age are summarized in

Tables 19, 20 and 21. Across the full age range small but

reliable trends are indicated for the SN preference scores

to increase with age for all samples having a dominant

perceptive attitude-function, with the exception of female

extraverted sensing types. Findings for the extraverted

sensing sample also indicate a positive association; how­

ever, the results are not statistically reliable (see Table

19). During the second developmental stage, the analyses of

trends summarized in

associations for all

types for both genders.

89

Table 20, show reliable positive

samples except extraverted sensing

For the third developmental stage,

the results show that for four samples (female and male

extraverted intuition, female introverted intuition, and

male extraverted sensing) a negative trend is indicated (see

Table 21). Although none of the findings for the third

stage are statistically reliable, the BETA values reported

are as large for some samples as those that are reported as

statistically reliable for the second stage. That is to say,

the trends of association are as strong, but the variance

estimates from the much smaller samples do not permit the

same degree of confidence in the findings.

Among the samples reporting a dominant perceptive

attitude-function, the research predictions for the second

developmental stage are supported for seven of eight

samples. For the remaining sample, female extraverted

sensing, the BETA value observed indicates a positive, but

unreliable, trend. The research predictions for the third

stage--of no reliable trends--are supported for all eight

groups; however, the magnitudes of the reported BETA values

suggests the possibility that larger samples, providing more

stable variance estimates, might not show the mix of

positive and negative trends to be reliable.

The generally close correspondence of the means for

the two genders seen in Figures 4 and 5, is strongly

90

suggestive that the age effects observed in tnese data

affect males and females in a similar manner. Also of

interest are the differing levels of reported preference by

samples in each age group. Among extraverts, intuition is

more consistently reported by intuitives than sensing is by

sensing types. Among introverts, the reverse relationship

is found: the sensing types report their sensing preference

more consistently than the intuitives report their

preference for intuition.

Auxiliary Sensing-Intuition

The predictions made for the preference scores

associated with the auxiliary attitude-functions were all

identical. During the second developmental stage the pre­

ference scores were expected to increase and the increase

was expected to be greater than when the preference score is

associated with a dominant attitude-function. During the

third stage, the preference scores were expected to continue

to increase with age, but more moderately. Therefore, it

was predicted that the SN preference scores would have

moderate and reliable increases with the respondents' age

during the second stage and small but reliable increases

with the respondents' age during the third stage.

The means of the SN preference scores for the eight

samples having a perceptive auxiliary function are plotted

in Figures 6 and 7. Considerable variability of the means

can be seen for each of the samples and, at least for

91

portions of the age ranges of each sample, an increase in

the preference score appears to be related to the increased

age of the groups.

The analyses of variance by age groups, summarized in

Table 22, confirm the statistically reliable differences

among the mean auxiliary SN preference scores for all

samples. However, Bartlett-Box tests indicate departures

from homogeneity of variance for three of the samples:

female extraverted sensing, female introverted intuition,

and male introverted sensing. Analyses of variance by

stages, summarized in Table 23, indicate that all samples

have reliably different means for the second and third

stages, except for females whose auxiliary is introverted

sensing and males whose auxiliary is introverted intuition.

However, in these analyses, departures from homogeneity of

variance also have been detected for three samples:

extraverted sensing (both genders ) and female introverted

sensing.

Regression analyses for trends of association between

auxiliary SN preference scores and age across the full age

range are summarized in Table 24. All trends show positive

association; however, the trends for female extraverted

sensing and both male and female extraverted intuition are

not statistically reliable. The gender difference among

the samples with auxiliary extraverted sensing is note­

worthy. The strongest degree of association observed in

92

these samples is among males, the weakest is among females.

This is in contrast to the gender similarities for the

introverted functions and for extraverted intuition.

Regression analyses for stage two, summarized in Table

25, showed that small, statistically reliable associations

between age and auxiliary SN preference scores are present

during this stage, except among the female auxiliary extra­

verted sensing sample. For the third developmental stage no

reliable trends are observed (see Table 26).

The results obtained from the analyses of SN

preference scores associated with the auxiliary atti tude­

functions demonstrate the trend to increase with age during

the second stage, as was predicted. Further, the absence of

reliable trends during the third stage is also consistent

with predictions. However, the negative trends found for

the third stage (even though they are not statistically

reliable) are not consistent with the theoretical model.

Dominant Thinking-Feeling

As noted above, I made parallel research predictions

for all of the attitude-functions. The second developmental

stage (age 15-39) is theoretically related to the greatest

period of development of the auxiliary attitude-function and

to a lesser extent to the further development of the

dominant attitude-function. During the third developmental

stage, the demands of

identification of ego

individuation require a lessening of

consciousness with the dominant

93

attitude-function, with an increasing interest in the pre-

viously undeveloped attitude-functions. Therefore, one

prediction was that the MBTI SN preference scores related to

the dominant attitude-function would show a small yet

reliable increase during the second developmental stage,

then remain constant during the third developmental stage.

In the analyses of the samples reporting a dominant

judging attitude-function (thinking or feeling), the TF

preference scores were predicted to increase with age during the second stage (age 15-39), and remain constant there-

after.

The plots of the mean TF preference scores are

displayed in Figures 8 and 9. Upon visual inspection, the

TF preference appears to increase with age only among the

male extraverted thinking sample. This observation is con-

firmed by analyses of variance, both by age groups and by

stage, which indicate reliable differences among the group

means only for the male extraverted thinking sample (see

Tables 27 and 28). However, the probability (2 = .0054) of

the result for the analysis of variance by age groups for

the female extraverted thinking sample barely fell short of

the criterion (2 = .005) for reliability in this study.

Regression analyses for trends of association across

the full age range indicate a reliable trend in the male

extraverted thinking sample for the TF preference score to

increase with age (see Table 29). During the second devel-

94

opmental stage both male and female extraverted samples

increase their preferences for thinking (see Table 30). No

reliable trends are observed for the third stage (see Table

31), and the largest of the nonreliable trends are negative.

The small increases in TF preference scores during the

second developmental stage (which were expected for all of

the samples reporting a preference for a dominant judging

function) appear in these data only among the samples pre­

ferring extraverted thinking. As expected, no reliable

trends were observed during the third stage, although, as

with other analyses reported here, the BETA values are large

enough (particularly for male introverts) to suggest that

reliable trends might be detected in larger samples.

Because the TF scale is scored with different weights

for males and females, the differences in level of

preference scores that appear in Figures 8 and 9 are likely

to reflect only the differences in the maximum available

scores. However, it is interesting to note the apparent

difference in level of TF preference for extraverted versus

introverted functions. Introverted feeling appears to be

reported more consistently at all age groups than is

extraverted feeling,

extraverted thinking

introverted thinking.

and older males appear to report

more consistently than they report

Although, these differences are

suggested by the plots of the means, they were not tested

statistically.

95

Auxiliary Thinking-Feeling

As with the dominant attitude-functions, the predic­

tions made for the preference scores associated with the

auxiliary attitude-functions are all identical. During the

second developmental stage the preference scores are

expected to increase; the increase is expected to be greater

than when the preference score is associated with a

dominant attitude-function. During the third stage, the

preference scores are expected to continue to increase with

age, but more moderately. Therefore, one prediction was

that the TF preference scores would have moderate, and

reliable associations with the respondents' age during the

second stage, and small but reliable associations with the

respondents' age during the third stage.

The plots of the mean auxiliary TF preference scores

are displayed in Figures 10 and 11. Visual inspection of

the plots suggests that there is somewhat more 'bounce' in

the pattern of variability for the judging preference when

it is the auxiliary function than when it is the dominant

(compare Figures 8 and 9 to Figures 10 and 11). Positive

trends appear most likely among the two male samples

reporting auxiliary thinking, though both samples show a

pronounced decline in the consistency of their preference

among the 50-to-59-year-old respondents (age group 7).

Analyses of variance by age groups, summarized in

Table 32, indicate reliable differences among the mean

96

auxiliary TF scores for the male and female extraverted

thinking samples and for the male introverted thinking

sample. Analyses of variance by developmental stages,

summarized in Table 33, show that the differences of the

means among the males are also present across the stages,

while the differences among the female extraverted thinking

sample are not.

Regression analyses across the full age range indicate

that small and reliable, positive trends associate age and

auxiliary TF preference scores for all of the introverted

functions and for male extraverted thinking (see Table 34).

During the second developmental stage, only the trend among

the male introverted thinking sample is reliable, although

the probability (2 = .0073) of the trend among the male

extraverted thinking sample closely approaches the criterion

for reliability (see Table 35). During the third stage

small negative, nonreliable, trends are present among

extraverted thinking samples of both sexes and the female

introverted thinking sample, and a small non-reliable trend

of positive association appears among the female

verted feeling sample.

intro-

In general, the pattern of trends for auxiliary TF

preference scores to increase with the age of the respondent

during the second developmental stage is not present in

these data. The trends observed during the third stage are

not statistically reliable, but suggest that both positive

97

and negative associations of age and preference scores may

occur.

Additional observations of the patterns of the

attitude-function preference scores have not been tested

for statistical reliability, yet are of note for theoret­

ical and methodological reasons. These observations concern

the utility of the distinction made in these analyses

between extraverted and introverted functions, and between

dominant and auxiliary functions. The value of the extra­

verted-versus-introverted distinction is most obvious with

respect to the relative levels of the preferences for

sensing and intuition (see Figures 4 and 5). Among

introverted sensing samples the preference for sensing is

noticeably more consistent than the intuitive preference of

the introverted intui ti ves. Conversely, the extraverted

intuitives reported intuition more consistently than the

extraverted sensing samples reported sensing. Though not to

the same extent, a similar difference of level

distinguishing extraverted feeling from introverted feeling

among females appears to be present.

The distinction between dominant and auxiliary

functions may also prove useful with respect to

understanding changes in the consistency of preference at

differing ages. The pattern of the means for extraverted

sensing among males appears to differ according to whether

the attitude-function is dominant or auxil iary (compare

98

Figures 4 and 7). Further, the homogeneity of the variance

of the preference scores is more consistent among analyses

in which the variance was associated with the dominant

rather than the auxiliary function. This is most clear for

the SN preference scores where 7 of 16 analyses showed

departures from homogeneity. That the dominant differs from

the auxiliary in this respect may reflect the generally less

even development of the auxiliary function that is expected

from theory.

Analyses of the Item Weights

Findings of statistically reliable changes in the

consistency of MBTI preference scores related to the ages of

the respondents prompted the present research. The pub­

lished literature contains several references to such

findings. My pilot analyses of a large unselected sample of

MBTI records substantially agreed with the published

reports. The methodological deficiencies of sample

selection in both the published analyses and the pilot

analyses provided one possible explanation for the reported

results. However, in reviewing the development of the MBTI

scales and the selection of the items comprising the

scales, I found a second possible explanation. The MBTI has

been revised 10 times since the first selection of the

items. New items and modifications of the scoring proce-

dures, such as differential weighting of the responses, were

introduced and revised to control for the relative

99

predictive value and social desirability of the responses.

Scoring procedures that had been adjusted initially to a

mature sample were revised for progressively younger samples

of respondents. The final item response weights were estab­

lished mainly using high school student samples. Thus, the

items and/or the assigned weights may be more appropriate

for younger, rather than older, respondents.

Myers assigned the item weights according to a predic­

tion ratio (PR) formula that may vary from zero to one.

Responses obtaining PR values of .630 to .729 received a

weight of one point, responses with a PR of .73 or above

received a weight of two points, and responses with PR

values below .630 or are endorsed by 50% or more respondents

having the opposite preference received a weight of zero

points.

In conducting the analyses of response weights in this

study I made the assumption that Myers's use of the

prediction ratios and response popularity were appropriate

methods for assessing the capacity of a response to predict

type preferences. Myers's standards were rigorous. Myers

(1961) reported that she discarded as inadequate thousands

of proposed item responses that failed to meet her

standards. Her early studies of items were based on adults

who were functioning well in society. Later she worked with

samples at lower ages and educational levels so that the

Indicator be able to predict type in these groups, where

100

type may be more difficult to assess. The current published

weights are based mainly on high school and younger student

samples.

The central question

data from the MBTI data

I addressed, using the

bank, was whether the

current

weights

established more than a decade ago with younger samples are

less than, equal to, or more than weights that would be

assigned by Myers's formulae on the basis of data bank

responses from older age groups. Responses where Myers's

weights are higher than weights that would be set by my

analyses of the data bank responses for any given age group

are designated over-weighted--that is, according to the

results of these analyses, the published scoring keys weight

the predictive value of the response higher than or over its

data bank weight. Responses where Myers's weights are

lower than weights established from data bank responses of

that age group are designated under-weighted--that is,

according to the results of these analyses, the published

scoring keys weight the predictive value of the response

lower than or under its data bank weight.

In the present study, to evaluate the appropriateness

of the item weights at different ages, I calculated the

prediction ratios for each of the scored items separately

for respondents at each age level used in the analyses of

variance and the regression analyses. Responses were

considered over-weighted if the calculated PR is less than

101

the minimum PR specified by Myers for the response's

assigned weight, or if the popularity of the response

exceeded 50&. Responses are considered under-weighted if

the calculated PR was greater than the maximum PR specified

by Myers for the response's assigned weight, and if the

popularity of the response does not exceed 50%. Responses

are considered correctly weighted if the calculated PR was

between the maximum and minimum PRs specified by Myers for

the response's assigned weight, and if the responses popu­

larity did not exceed 50%.

Where responses were under-weighted, I calculated the

sum of the potential under-weighting of the point totals for

that preference. Likewise, where responses were over­

weighted, I calculated the sum of the potential over­

weighting of the point totals for that preference. For each

preference, I calculated a total potential over-weighting

and a total potential under-weighting. The results of these

calculations appear in Tables 37-40 for the EI, SN, TF, and

JP scales, respectively.

These tables require some detailed explanation. In

Table 37 the calculations for the E and the I responses for

the age groups are displayed by rows. The separate sums are

arranged in blocks of three columns for female, male, and

combined sexes, respectively. The first two blocks refer

only to responses weighted for

second two blocks refer only to

extraversion, while

responses weighted

the

for

102

introversion. The first block displays the sums obtained by

adding, for all responses with an extraverted weight, the

point deviations obtained by comparing the published scoring

weights (Myers, 1977) to the weights prescribed by the

current prediction ratios calculated from the sample of MBTI

records used in this study. The point sums in the first

block are from responses that are under-weighted according

the current prediction ratios. The point sums in the second

block are from responses that are over-weighted according to

current prediction ratios. The third and fourth blocks are

from introverted responses that are under-weighted or over­

weighted, respectively. The sums are for points rather than

counts of items because the two-point range of the possible

weights permits a two-point weight to fall to zero, or a

zero weight to rise

would underestimate

to two points.

the potential

Counts of responses

effects of the point

deviations on the resulting preference scores. In each row,

data for the age group are presented in the order of

females, males, and total sample to facilitate comparisons

of gender differences.

For example, the last row of Table 37 shows the point

sums calculated for all ages combined and the last column of

each block shows the point sums for both sexes combined.

The point deviation sums for all ages and gender combined

are the sums that most directly relate to the published

scoring weights because the published keys for the EI scale

103

do not distinguish between gender or age. Three extraverted

responses were found to be under-weighted by one point each,

nine extraverted responses were found to be over-weighted by

one point each, six introverted responses were found to be

under-weighted by one point each, and four introverted

responses were found to be over-weighted by one point each,

while another introverted response was found to be over­

weighted by two points. The corresponding sums, then, are

3, 9, 6 , and 5.

No simple method is available to determine the actual

effects on the resulting preference scores that a particular

set of deviations from the published weights potentially

may have. For some items, one response may be correctly

weighted and its alternative response is under- or over­

weighted. In that instance the deviation would affect only

the scores for respondents who selected the under-weighted

response. Some patterns of responses to the scale items

could result in none of the improprely weighted responses

being selected. other patterns might neatly balance the

selection of under-weighted and over-weighted responses to

produce no net effect. Finally, some patterns might, as

though perversely guided, select responses improperly

weighted in one direction only, thereby maximizing the

distortion of the resulting preference score in that

direction.

104

We can only examine relative size of the sums at the

different age levels to speculate on the likely direction of

the general effect. In so speculating, it is necessary to

realize that selecting under-weighted responses for one

preference produces the same net effect on the preference

scores as does selecting over-weighted responses for the

opposite preference (see page 28 above for a description of

the MBTI scoring procedures). Specifically, under-weighted

extraverted responses and over-weighted introverted

responses combine to underestimate the extraverted

respondent's consistency of the preference for extraversion,

or over-estimate the consistency of the introvert's

preference for introversion. Further, the effects of the

point differences on the preferences is doubled due to the

tie-breaking formula that doubles the difference between

the point totals to calculate the preference score (see page

29) •

Therefore, examining the sums from the EI scale

example above for all ages and both sexes combined, we find

extraversion under-weighted by 3 points, introversion over­

weighted by 5 points, extraversion over-weighted by 9

points, and introversion under-weighted by 6 points. The

combined effects moving preference scores toward

introversion and away from extraversion are larger than the

opposi te effects (9 and 6 versus 3 and 5). Al though the

extent of the effect will vary with the actual pattern of

105

responses selected by any given respondent, the general

effect in this instance is for the preference for

introversion to be underestimated, or the preference for

extraversion to be overestimated.

This finding has two consequences. First, if the MBTI

records were to be rescored with the weights prescribed by

the present analysis, there would be a tendency for all

respondents to obtain scores indicating more introversion.

Second, some respondents with low scores would be re­

classified from extraverts to introverts. The scores from

these cases have been included in analyses of age trends for

extraverts (reported above), when, if rescored, they would

have been included in the analyses of trends for introverts.

The finding of improperly weighted responses based on

the total sample was not expected in this study. Never-

theless, based on the type of analysis just presented for

the EI scale, Table 38 shows that a small overall effect is

also found for overestimating intuition and underestimating

sensing (3 and 3 versus 2 and 1). The TF scale is weighted

separately for males and females, thus the effects must be

considered separately. In Table 39, no effect occurs for

males, but for females a strong effect is found to result in

overestimation of feeling and underestimation of thinking

(3 and 5 versus 0 and 0).

Finally, as shown in Table 40, a strong effect toward

overestimating judgment and underestimating perception is

106

found (1 and 5 versus 6 and 6). Calculating the point

deviation sums for the JP responses is complicated by items

having three or more responses. The sums in parentheses in

Table 40 result from correctly weighted and improperly

weighted responses for the same preference being available

for a particular item. The parenthetical entries are for

sums calculated based on the responses with the least

deviation from the published weights. The effect to over­

estimate judgment and underestimate perception is very

similar if those sums are used (0 and 4 versus 6 and 4).

These results raise issues I have not previously

addressed in this investigation. The predictive value and/or

the popularity of the item responses scored for

psychological type has changed markedly from the analyses

Myers (1975) conducted for the El, SN, and JP scales in

preparation for the publication of the MBTl Form F in 1962;

for females these parameters have changed from Myers (1977)

analysis of the TF scale. Myers (1977) argued that the

revision was necessary for the TF scale because of

generation differences in the social desirability of the

thinking and feeling responses respective to the changing

social roles of men and women. The results of this study

indicate that the changes extend beyond the thinking and

feeling responses and that each of the scales need revised

item weights.

107

However, the possibility of a shift in response

weights for all respondents does not account for the

associations of the preference scores with age in a cross­

sectional sample. All of these data were collected during

the same period, 1978 to 1984. The purpose of these

analyses was to determine if

equally appropriate for all

the

the

response weights are

age groups sampled.

Therefore, differences of overestimation or underestimation

of a preference score at different ages is the particular

interest of this study.

Examining the sums of the point deviations for

different age groups discloses that the deviations are not

not constant across the age groups. A tendency for increase

appears in some columns, and for decrease in others. For

some preferences, the pattern appears curvilinear.

There is no precise way to determine the effects which

the age differences of the sums of the point deviations

have on the trends of association reported earlier, without

actually rescoring the CAPT data bank and replicating the

analyses reported above. However, the trends may be

expected to be altered if the records would be scored by

revised weights.

The tables just presented are useful in speculating

about the effects of improper weighting of the item

responses, but do not provide statistically reliable tests

of the associations of the selection of particular item

108

responses with the respondents' age. To supplement the

foregoing analyses, I performed additional analyses that

statistically test the reliability of the associations. To

achieve this, I tabulated for each item contingency tables

of the item responses for each sex separately by age

groups. If no association exists between age group and

response selection, nonreliable chi-square tests would be

obtained for each contingency table.

Among both sexes, 19 of 22 EI items are reliably

associated with the respondents' age at a probability of .05

or less. Among females 20 of 26 SN items and among males 20

of 26 SN items are reliably associated with the respondents'

age (though it was not the same 20 items for both sexes).

Among females 17 of 23 TF items and among males 21 of 23 TF

items are reliably associated with the respondents' age.

For both sexes, 21 of 24 JP items are reliably associated

with the respondents' age.

Of all the 95 items scored for type, 81% are

associated with age for females, and 85% are associated with

age for males.

35 3-4 33 32

P 31 R30 E29 F~

109

Extraversion­Introversion Scores

Total Sample N = 21,696

M E 26 E R 25 A E 2-4 1----__

N N 23 ............... -- .. '

C22 E E 21

I S ~ C 18 o 17 R 16 E 15

1-4 13

...... ..... ... ........... . . . ~-•• / .. - • .,...c=:n. ___ ~..L-:-: • •• ' ••• ....t...&.-.. • ••••• , • '_If .. _ ------. .---... ........ \,...,...........~. . .--......... ~ ... ~ .. -~ ..... ~- .~,

12 1l+-----~----~~----4------+------~----~-----4

1 2 3 4 5 6 7 B Age Group

- Female ... Female .- Male ... Male Extraverts Introverts Extraverts Introverts

Figure 1. Mean EI Preference Scores by Age Groups

110

Table 7. Analyses of Variance: EI by Age Groups

Gender Attitude F ratio Probability

Female Extraversion 13.88 .0000 a

Female Introversion 4.68 .0000

Male Extraversion 2.94 .0044

Male Introversion 3.59 .0007

a. Bartlett-Box test of homogeneity of variance indicates departure from homogeneity at probability < .05.

Table 8. Analyses of Variance: EI by Developmental stages

Gender Attitude F ratio Probability

Female Extraversion 10.14 .0015

Female Introversion 2.62 .1056

Male Extraversion 1. 27 .2601

Male Introversion 2.73 .0988

Table 9. Regressions of EI by Age: All Ages

Gender Attitude BETA F ratio Probability

Female Extraversion -.1045 59.84 .0000

Female Introversion .0444 10.71 .0011

Male Extraversion .0202 2.21 .1368

Male Introversion .0397 8.58 .0034

III

Table 10. Regressions of EI by Age: stage Two

Gender Attitude BETA F ratio Probability

Female Extraversion -.1312 83.35 .0000

Female Introversion .0490 11.43 .0007

Male Extraversion .0172 1.41 .2357

Male Introversion .0528 13.33 .0003

Table 11. Regressions of EI by Age: stage Three

Gender Attitude BETA F ratio Probability

Female Extraversion -.0347 0.80 .3818

Female Introversion .0245 0.40 .5284

Male Extraversion -.0120 0.10 .7569

Male Introversion -.0592 2.33 .1278

35 34 33 32

P 31 R30 E29 F~

M E 26 E R 2S " E 24 N N 23

C22 J E 21

P 5 ~ C 18 o 17 R 16 E 15

14 13

112

Judging-Perceptive S-cores:

Extraverts N = 10,848

". . ... . . ..... . . / ---. - ... _------_:._,.,-----,. .... ". ~:.::. .... -----""""'"

" . . . . -------------- .. , . " -----. . ". " . ' .. " . , . , . , .... '. ~, ., ..

12 11+------4------~------r------+------~----_4~----~

I 2

- Female Extraverted Rational

3 4 5 Age Group

--- Female Extraverted Irrational

.- Male Extraverted Rational

6 7

... Male Extraverted Irrational

Figure 2. Mean JP Preference Scores by Age Groups: Extraverts

B

3S 34 33 32

P 31 R30 E~ F"

M E 26 E R 25 A E 24 N N 23

C22 J E 21

P S ~ C 18 o 17 R 16 E 15

14 13

113

Judging-Perceptive Scores

Introverts N = 10,848

.. to .... '_ :7' " . . . . . . . . .._,..------------------...... 7 - --------- .. -..... .. :,..:....... .. .. ...., ..... ----------- ...... . --. . ........ .......... .. ":'., ... ....

12 11+------+------+------+------+------+------+-----~

1 2

- Female Introverted Irrational

3 4 5 AgQ Group

--. Female Introverted Rational

.- Male Introverted Irrational

6 7

... Male Introverted Rational

Figure 3. Mean JP Preference Scores by Age Groups: Introverts

8

114

Table 12. Analyses of Variance: JP by Age Groups

Gender Attitude Scale F ratio Probability

Female Extraverted Rational J 8.89 .0000

Female Extraverted Irrational P 3.78 .0004

Female Introverted Rational P 0.97 .4539

Female Introverted Irrational J 9.27 .0000

Male Extraverted Rational J 28.21 .0000

Male Extraverted Irrational P 5.10 .0000

Male Introverted Rational P 6.05 .0000

Male Introverted Irrational J 17.50 .0000

Table 13. Analyses of Variance: JP by Developmental Stages

Gender Attitude Scale F ratio Probability

Female Extraverted Rational J 20.63 .0000

Female Extraverted Irrational P 2.24 .1350

Female Introverted Rational P 0.85 .3547

Female Introverted Irrational J 26.62 .0000

Male Extraverted Rational J 63.97 .0000

Male Extraverted Irrational P 6.60 .0102

Male Introverted Rational P 12.28 .0005 a

Male Introverted Irrational J 46.84 .0000

a. Bartlett-Box test of homogeneity of variance indicates departure from homogeneity at probability < .05.

115

Table 14. Regressions of JP by Age: All Ages

Gender Attitude Scale BETA F ratio Prob.

Female Extraverted Rational J .1362 51. 25 .0000

Female Extraverted Irrational P -.0617 10.36 .0013

Female Introverted Rational P .0000 0.00 .9861

Female Introverted Irrational J .1406 54.66 .0000

Male Extraverted Rational J .2304 151. 96 .0000

Male Extraverted Irrational P -.0884 21. 33 .0000

Male Introverted Rational P -.1006 7.65 .0000

Male Introverted Irrational J .1917 103.42 .0000

Table 15. Regressions of JP by Age: Stage Two

Gender Attitude Scale BETA F ratio Prob.

Female Extraverted Rational J .1089 28.54 .0000

Female Extraverted Irrational P -.0647 10.00 .0015

Female Introverted Rational P -.0154 0.56 .4526

Female Introverted Irrational J .1145 31. 58 .0000

Male Extraverted Rational J .2052 104.48 .0000

Male Extraverted Irrational P -.0743 13.20 .0003

Male Introverted Rational P -.0934 21. 03 .0000

Male Introverted Irrational J .1683 69.31 .0000

116

Table 16. Regressions of JP by Age: stage Three

Gender Attitude Scale BETA F ratio Prob.

Female Extraverted Rational J .1283 5.53 .0193

Female Extraverted Irrational P -.0657 1. 43 .2322

Female Introverted Rational P -.0724 1. 74 .1884

Female Introverted Irrational J .0002 0.00 .9757

Male Extraverted Rational J .0625 1. 29 .2564

Male Extraverted Irrational P -.1389 6.49 .0113

Male Introverted Rational P .0221 0.16 .6885

Male Introverted Irrational J .0278 0.25 .6142

,

3S 34 n 32

p 31 R30 E29 F~

M E 26 E R 2S A E 24 N N 23

[22 S E 21

N S ~ C 18 017 R 16 E 15

14 13

117

Sensing-Intuition Scores Extraverts N = 10,848

.. ...... ............ . ' '" .. ~~;- .. ------------....... f"l-t--- ----_ '. ........ ;/ . " '. . , " .

", " ' . ", " '. ....... ' . ", /.,.,--,--.

"'~ '.,.. ~ ......... .' ."", /' ........ :..:. .... " ~.--.----.--" --.---. . _.--. -------

12 11+------4------4------4~----~------r_----_r----~

1 2

- Female Extraverted Sensing

3 4 5 Age Group

--. Female Extraverted Intuition

.- Male Extravertod Sensing

6 7

. .. Male Extraverted Intuition

Figure 4. Mean SN Preference Scores by Age Groups: Extraverts

B

118

SQnsing-Intuition ScorQS IntrovQrts N = 10,848

35.55 35.57 3S 34 33 32

P 31 R30 E29 F~

M E 26 E R 2S A E 24 N N 23

C22 S E 21

N S ~ C 18 o 17 R 16 E 15

14 13

..... ' ..

- • I •••••••• ••• .. • ........ ..

",,,, -.. _---- ..... -.... " ...

.. .. .. . .. .. ,,,,,' ----------.. .. .' .. , ... ' ---------',~.... ...... ,

............... ~ .. I. .. .. .. .._---'

' .... _---------------

12 11+------+------+------+------+------+------+-----~

1 2

- Female Introverted Sansing

3 4 5 Age Group

--- Female Introverted Intuition

'- Male Introverted Sensing

6 7

... Male Introverted Intuition

Figure 5. Mean SN Preference Scores by Age Groups: Introverts

B

119

Table 17. Analyses of Variance: SN by Age Groups

Gender Attitude-Function F ratio Probability

Female Extraverted Sensing 1. 96 .0569

Female Extraverted Intuition 8.17 .0000

Female Introverted Sensing 3.13 .0027 a

Female Introverted Intuition 3.63 .0007

Male Extraverted Sensing 2.51 .0143

Male Extraverted Intuition 8.63 .0000

Male Introverted Sensing 6.49 .0000 a

Male Introverted Intuition 3.95 .0003

a. Bartlett-Box test of homogeneity of variance indicates departure from homogeneity at probability < .05.

Table 18. Analyses of Variance: SN by Developmental Stages

Gender Attitude-Function F ratio Probability

Female Extraverted Sensing 6.42 .0114

Female Extraverted Intuition 7.30 .0070

Female Introverted Sensing 10.22 .0014

Female Introverted Intuition 3.64 .0565

Male Extraverted Sensing 6.49 .0110

Male Extraverted Intuition 10.24 .0014

Male Introverted Sensing 23.76 .0000

Male Introverted Intuition 3.28 .0703

120

Table 19. Regressions of SN by Age: All Ages

Gender Attitude-Function BETA F ratio Prob.

Female Extraverted Sensing .0932 11.84 .0931

Female Extraverted Intuition .1298 23.21 .0000

Female Introverted Sensing .1145 17.98 .0000

Female Introverted Intuition .0899 11.04 .0009

Male Extraverted Sensing .0762 7.90 .0050

Male Extraverted Intuition .1340 24.78 .0000

Male Introverted Sensing .1758 43.18 .0000

Male Introverted Intuition .1056 15.27 .0001

Table 20. Regressions of SN by Age: Stage Two

Gender Attitude-Function BETA F ratio Prob.

Female Extraverted Sensing .0562 3.76 .0526

Female Extraverted Intuition .1678 34.41 .0000

Female Introverted Sensing .0858 8.80 .0031

Female Introverted Intuition .1068 13.70 .0002

Male Extraverted Sensing .0665 5.28 .0217

Male Extraverted Intuition .1487 26.86 .0000

Male Introverted Sensing .1272 19.52 .0000

Male Introverted Intuition .1283 19.90 .0000

121

Table 21. Regressions of SN by Age: stage Three

Gender Attitude-Function BETA F ratio Probe

Female Extraverted Sensing .1016 1. 71 .1926

Female Extraverted Intuition -.1443 3.75 .0544

Female Introverted Sensing .0262 0.11 .7375

Female Introverted Intuition -.0488 0.39 .5322

Male Extraverted Sensing -.1572 4.16 .0413

Male Extraverted Intuition -.1338 2.99 .0856

Male Introverted Sensing .0766 0.97 .3265

Male Introverted Intuition .0003 0.00 .9684

35 34 33 32

P 31 R30 [29

F~ M E 26 E R 25 It E 24 N N 23

C22 S E 21

N S ~ C 18 o 17 R 16 E 15

14 13

122

Auxiliary Sensing-Intuition Scores

Introverts N = 10,848

. . -....... ./; .';--..

~~--~-- ------ _ .. ' .* ...... . ..:..:.~ •• v...... ..... ....... -----.. , .. ~

'--'-'.~~' .. " ..... / ., ~ ...

...... . ..' . . .... ~:. .. ~-~ ..... .... "-' ... .. .., ..

.......... : ........... .,'/. ....... ..... .. .. '

........... ---'

12 J1+-----_+------+-----_+------~----_+------~----~

1 2

- Fgmalg Extravgrtgd Sensing

3 4 5 "gg Group

--- Female Extravertgd Intuition

.- Male Extravertgd Sensing

6 7

... Malg Extravgrtgd Intuition

8

Figure 6. Mean Auxiliary SN Preference Scores by Age Groups: Introverts

35 3-4 33 32

P 31 R 30 E29 F~

123

Auxiliary Sensing-Intuition Scores Extraverts N = 10,848

/--:::.

/' ."..

.~.-.--. /

M E 26 E R 2S A E 2-4 N N 23 e22 ~. .. ...................... ...

E L--- .. .,._........ ...., ...... 5 21 .' .. ........... - . . . . . . . . . . . .. .. N 5 20

19 _#""........ . ' . . ... ~':<

,,;- . " C 18 .. ".... " o 17 . . ...... "

,. .. --...... ~~::.:.~ .. ------.." '" R 16 , E 15 '

1-4 13 12 11+------;------~------r------;-------+1------41------~1 123 4 5 6 7 B

- Famala Introvartad Sansing

Aga Group

... Famala Introvartad Intuition

.- Mala Introvartad Sansing

... Mala Introvartad Intuition

Figure 7. Mean Auxiliary SN Preference Scores by Age Groups: Extraverts

124

Table 22. Analyses of Variance: Auxiliary SN by Age Groups

Gender Attitude-Function F ratio Probability

Female Extraverted Sensing 4.92 .0000 a

Female Extraverted Intuition 7.48 .0000

Female Introverted Sensing 4.55 .0000

Female Introverted Intuition 6.09 .0000 a

Male Extraverted Sensing 8.49 .0000

Male Extraverted Intuition 5.84 .0000

Male Introverted Sensing 5.14 .0000 a

Male Introverted Intuition 4.67 .0000

a. Bartlett-Box test of homogeneity of variance indicates departure from homogeneity at probability < .05.

Table 23. Analyses of Variance: Auxiliary SN by Developmental Stages

Gender Attitude-Function F ratio Probability

Female Extraverted Sensing 20.30 .0000 a

Female Extraverted Intuition 22.32 .0000

Female Introverted Sensing 4.53 .0335 a

Female Introverted Intuition 12.99 .0003

Male Extraverted Sensing 32.93 .0000 a

Male Extraverted Intuition 9.96 .0016

Male Introverted Sensing 17.00 .0000

Male Introverted Intuition 5.37 .0206

a. Bartlett-Box test of homogeneity of variance indicates departure from homogeneity at probability < .05.

125

Table 24. Regressions of Auxiliary SN by Age: All Ages

Gender Attitude-Function BETA F ratio Prob.

Female Extraverted Sensing .0361 1. 77 .1841

Female Extraverted Intuition .0468 2.98 .0847

Female Introverted Sensing .1137 17.72 .0000

Female Introverted Intuition .1286 22.78 .0000

Male Extraverted Sensing .1994 45.05 .0000

Male Extraverted Intuition .0637 5.52 .0189

Male Introverted Sensing .1459 29.45 .0000

Male Introverted Intuition .1162 18.54 .0000

Table 25. Regressions of Auxiliary SN by Age: Stage Two

Gender Attitude-Function BETA F ratio Prob.

Female Extraverted Sensing .0307 1.12 .2896

Female Extraverted Intuition .1443 25.26 .0000

Female Introverted Sensing .1052 13.29 .0003

Female Introverted Intuition .1370 22.72 .0000

Male Extraverted Sensing .1410 24.08 .0000

Male Extraverted Intuition .1369 22.68 .0000

Male Introverted Sensing .1080 14.03 .0002

Male Introverted Intuition .1176 16.66 .0000

126

Table 26. Regressions of Auxiliary SN by Age: stage Three

Gender Attitude-Function BETA F ratio Probe

Female Extraverted Sensing .0772 0.98 .3231

Female Extraverted Intuition -.0193 0.06 .8048

Female Introverted Sensing .1524 3.90 .0500

Female Introverted Intuition -.2003 6.85 .0097

Male Extraverted Sensing .0688 0.78 .3784

Male Extraverted Intuition .0102 0.01 .8958

Male Introverted sensing .0292 0.14 .7092

Male Introverted Intuition .0681 0.76 .3834

35 34 33 32

P 31 R30 £29 F~

M E 26 E R 2S A E 24 N N 23

(22 T E 21

F S ~ ( 18 o 17 R 16 E 15

14 13

.-.~.

127

Thinking-Feeling Scores: Extraverts N = 10,848

~.---. --' ---.~

.~.--.--. .'" / .'" /. .

... ....... -..... -------,

" " " " " "

f"" " "" •

" " .................................... ., .. f 'f ............... • ' "

" 12 11+------+------+------+------+------+------r-----~

1 2

- Fgmole Extroverted Thinking

3 4 5 Age Group

--- Female Extroverted Feeling

.- Mole Extroverted Thinking

6 7

.. , Mole Extroverted Feel ing

Figure 8. Mean TF Preference Scores by Age Groups: Extraverts

B

35 34 33 32

P 31 R30 Era F"

M E 26 E R 2S A E 24 N N 23

C22 T E 21 F 20

S 19 C 18 o 17 R 16 E 15

14 13

10. ' .···,·,· '

128

Thinking-Faaling Scores: Introvarts N = 10,848

.. • ••••••• 10 .... .. ••

12 11+------+------~-----+------~----_r------r_----~

1 2

- Female Introvartad Thinking

3 4 5 Aga Group

--- Famala Introvartad FaQling

.- Male Introvertad Thinking

6 7

... Male Introvertad Faal ing

Figure 9. Mean TF Preference Scores by Age Groups: Introverts

8

129

Table 27. Analyses of Variance: TF by Age Groups

Gender Attitude-Function F ratio Probability

Female Extraverted Thinking 2.89 .0054

Female Extraverted Feeling 1. 68 .1091

Female Introverted Thinking 1. 35 .2213

Female Introverted Feeling 1.29 .2502

Male Extraverted Thinking 10.77 .0000

Male Extraverted Feeling 0.35 .9300

Male Introverted Thinking 1.15 .3310

Male Introverted Feeling 1. 62 .1251

Table 28. Analyses of Variance: TF by Developmental stages

Gender Attitude-Function F ratio Probability

Female Extraverted Thinking 0.01 .9034

Female Extraverted Feeling 1. 37 .2420

Female Introverted Thinking 0.56 .4550

Female Introverted Feeling 6.63 .0101

Male Extraverted Thinking 21. 38 .0000

Male Extraverted Feeling 0.18 .6707

Male Introverted Thinking 0.92 .3387

Male Introverted Feeling 0.40 .5268

130

Table 29. Regressions of TF by Age: All Ages

Gender Attitude-Function BETA F ratio Probe

Female Extraverted Thinking .0483 3.17 .0755

Female Extraverted Feeling .0007 0.06 .8060

Female Introverted Thinking .0408 2.26 .1329

Female Introverted Feeling .0529 3.80 .0516

Male Extraverted Thinking .1942 53.08 .0000

Male Extraverted Feeling .0007 0.62 .8060

Male Introverted Thinking .0275 1. 03 .3112

Male Introverted Feeling .0396 2.12 .1453

Table 30. Regressions of TF by Age: stage Two

Gender Attitude-Function BETA F ratio Probe

Female Extraverted Thinking .1058 13.44 .0003

Female Extraverted Feeling .0266 0.84 .3600

Female Introverted Thinking .0636 4.83 .0282

Female Introverted Feeling -.0126 0.19 .6639

Male Extraverted Thinking .2008 49.92 .0000

Male Extraverted Feeling .0297 1. 05 .3060

Male Introverted Thinking .0355 1. 50 .2208

Male Introverted Feeling .0281 0.94 .3313

131

Table 31. Regressions of TF by Age: stage Three

Gender Attitude-Function BETA F ratio Probe

Female Extraverted Thinking -.0908 1. 36 .2446

Female Extraverted Feeling -.1908 6.19 .0138

Female Introverted Thinking -.0892 1. 32 .2531

Female Introverted Feeling .0368 0.22 .6382

Male Extraverted Thinking -.0715 0.84 .3600

Male Extraverted Feeling .0124 0.03 .8742

Male Introverted Thinking -.1439 3.47 .0644

Male Introverted Feeling -.1341 3.00 .0849

35 34 33 32

p 31 R30 E~ F£1

M E 26 E R 25 A E 24 N N 23 Czz T E 21 F 20

5 19 C IB o 17 .. R 16 E 15

14 13

132

Auxiliary Thinking-Feeling Scores Introverts N = 10,848

.. .. .. .. .. .... .... .. .. .. ........ .. .. .... .. .. .... .. .. .... .. .... ... ................ " .. .......... ..

... .... .....

,.,.----------

12 11+------r------~----~----~------+_----_r---~ 123 4 5 6 7 8

- Fgmole Extroverted Thinking

Age Group

--- Female Extrovgrted Fegling

.- Malg Extrovgrtgd Thinking

... Mole Extroverted Feel ing

Figure 10. Mean Auxiliary TF Preference Scores by Age Groups: Introverts

35 34 33 32

P 31 R30 E29 F~

M E 26 E R 2S A E 24 N N 23

C22 T E 21

F S ~ C 18 o 17 R 16 E 15

14 f+-'-'--' ' -' -~ 13

133

Auxiliary Thinking-Feeling Scores Extraverts N = 10,848

." .. 12 7.71 11~-----r----~~----1------+------r-----~--~~

1 2

- Femah~ Introverted Thinking

3 4 5 Age Group

--- Female Introverted FeQl1ng

.- Male Introverted Thinking

6 7

... Male Introverted Feeling

Figure 11. Mean Auxiliary TF Preference Scores by Age Groups: Extraverts

8

134

Table 32. Analyses of Variance: Auxiliary TF by Age Groups

Gender Attitude-Function F ratio Probability

Female Extraverted Thinking 3.19 .0024

Female Extraverted Feeling 1. 58 .1381

Female Introverted Thinking 2.45 .0170

Female Introverted Feeling 1.19 .3037

Male Extraverted Thinking 4.02 .0002

Male Extraverted Feeling 0.96 .4578

Male Introverted Thinking 9.22 .0000

Male Introverted Feeling 2.00 .0514

Table 33. Analyses of Variance: Auxiliary TF by Developmental stages

Gender Attitude-Function F ratio Probability

Female Extraverted Thinking 2.13 .1443

Female Extraverted Feeling 0.20 .6583

Female Introverted Thinking 1. 61 .2045

Female Introverted Feeling 3.11 .0780

Male Extraverted Thinking 12.01 .0005

Male Extraverted Feeling 1.49 .2219

Male Introverted Thinking 18.21 .0000

Male Introverted Feeling 5.18 .0230 a

a. Bartlett-Box test of homogeneity of variance indicates departure from homogeneity at probability < .05.

135

Table 34. Regressions of Auxiliary TF by Age: Age Groups

Gender Attitude-Function BETA F ratio Prob.

Female Extraverted Thinking .0346 1. 63 .2028

Female Extraverted Feeling -.0189 0.49 .4862

Female Introverted Thinking .1210 20.11 .0000

Female Introverted Feeling .1732 41.90 .0000

Male Extraverted Thinking .0980 13.13 .0003

Male Extraverted Feeling .0270 0.98 .3213

Male Introverted Thinking .2000 56.41 .0000

Male Introverted Feeling .1400 27.05 .0000

Table 35. Regressions of Auxiliary TF by Age: stage Two

Gender Attitude-Function BETA F ratio Prob.

Female Extraverted Thinking .0368 1. 79 .1807

Female Extraverted Feeling -.0216 0.55 .4572

Female Introverted Thinking .0458 2.50 .1140

Female Introverted Feeling -.0128 0.19 .6599

Male Extraverted Thinking .0777 7.22 .0073

Male Extraverted Feeling .0007 0.05 .8210

Male Introverted Thinking .1676 34.33 .0000

Male Introverted Feeling .0245 0.72 .3977

136

Table 36. Regressions of Auxiliary TF by Age: stage Three

Gender Attitude-Function BETA F ratio Probe

Female Extraverted Thinking -.1774 5.33 .0223

Female Extraverted Feeling .0193 0.05 .8051

Female Introverted Thinking -.1482 3.68 .0567

Female Introverted Feeling .1684 4.79 .0301

Male Extraverted Thinking -.1853 5.83 .0168

Male Extraverted Feeling -.0377 0.23 .6300

Male Introverted Thinking .0264 0.11 .7360

Male Introverted Feeling .0157 0.02 .8407

137

Table 37. Point Deviations By Age Groups: EI Items

Sum of E Points Sum of I Points

Under- Over- Under- Over-weighted weighted weighted weighted

GENDER: F M Tot F M Tot F M Tot F M Tot

Age

15 - 17 4 4 4 7 7 8 7 6 7 3 6 5

18 - 20 3 3 3 8 8 8 7 5 7 5 7 5

21 - 24 4 4 3 6 9 9 7 5 6 5 6 5

25 - 29 5 3 3 6 9 7 5 4 6 7 7 7

30 - 39 5 4 4 6 6 6 5 5 5 8 5 6

40 - 49 3 4 4 6 7 6 5 6 6 7 5 6

50 - 59 4 5 5 6 8 7 4 5 5 7 6 6

60 & over 5 5 5 7 9 6 5 5 4 7 5 5

15 - 39 4 3 3 7 7 8 7 5 6 6 5 5

40 & over 4 4 4 6 6 6 5 6 6 7 6 6

All ages 4 3 3 6 7 9 7 5 6 4 5 5

138

Table 38. Point Deviations By Age Groups: SN Items

Sum of S Points Sum of N Points

Under- Over- Under- Over-weighted weighted weighted weighted

GENDER: F M Tot F M Tot F M Tot F M Tot

Age

15 - 17 3 2 2 2 2 1 1 1 1 9 5 7

18 - 20 2 1 2 3 3 3 3 1 1 7 6 6

21 - 24 6 6 5 4 2 2 1 2 1 4 5 5

25 - 29 7 6 6 2 2 2 2 4 3 4 5 5

30 - 39 9 10 12 1 2 1 5 7 8 3 3 2

40 - 49 9 12 11 1 1 1 7 8 8 3 2 2

50 - 59 9 7 10 1 0 0 7 9 9 2 3 2

60 & over 9 7 9 3 2 2 8 9 9 3 5 5

15 - 39 3 3 3 3 2 2 2 2 1 3 4 4

40 & over 10 11 8 1 1 1 7 8 8 2 2 2

all ages 5 3 3 3 1 2 2 4 1 3 3 3

139

Table 39. Point Deviations By Age Groups: TF Items

Sum of T Points Sum of F Points

Under- 1 Over- Under- Over-weighted 1 weighted weighted weighted

1-GENDER: F M 1 F M F M F M

1 Age 1

1 15 - 17 2 0 1 2 0 0 0 8 3

1 18 - 20 2 0 1 1 0 0 0 6 3

1 21 - 24 2 1 0 0 0 0 3 2

25 - 29 2 2 0 1 0 2 3 1

30 - 39 4 5 1 0 0 2 3 0

40 - 49 4 3 1 0 1 4 2 0

50 - 59 6 5 1 2 1 3 6 2

60 & over 3 7 5 2 1 2 9 2

15 - 39 3 0 0 1 0 0 4 1

40 & over 5 5 0 0 1 3 5 0

All ages 3 1 0 0 0 0 5 0

140

Table 40. Point Deviations By Age Groups: JP Items

Sum of J Points Sum of P Points

Under- Over- Under- Over-weighted weighted weighted weighted

GENDER: F M Tot F M Tot F M Tot F M Tot

Age

15 - 17 1 3 2 5 3 2 5 4 5 7 5 6 (0) (2) (1) (4) (2) (4) (6) (4) (5)

18 - 20 2 2 1 5 5 5 7 4 5 7 5 6 (1) (1) (0) (4) (2) (5) (6) (4) (5)

21 - 24 3 2 1 6 5 5 9 6 8 6 4 5 (1) (0) (6) (4) (6) (5) (3) (4)

25 - 29 0 1 1 5 6 5 9 7 6 5 4 4 (0) (0) (6) (4) (4) (4) (3) (3)

30 - 39 2 3 1 5 5 4 9 8 7 3 2 2 (0) (2) (0) (7) (5) (5) (2) (1) (1)

40 - 49 2 1 2 7 4 5 8 8 9 3 1 2 (1) (0) (1) (6) (5) (6) (3) (1) (2 )

50 - 59 2 2 2 3 6 5 9 5 6 1 3 3 (1) (1) (1) (7) (2) (3 ) (2) (2)

60 & over 4 1 1 5 7 5 12 9 8 I 3 1 3 (2) (0) (0) (9) (6) (5) I (0) (2)

I 15 - 39 1 2 1 6 6 6 8 4 4 I 6 4 5

(0) (1) (0) (6) (2) (3) I (5) (3 ) (4) I

40 & over 2 0 2 6 5 6 I 11 5 9 I 3 2 3 (1) (1) I (8) (2) (6) I (1) (2)

I I All ages 1 1 1 7 5 6 I 8 5 6 I 5 4 5

(0) (0) (0) I (6) (2) (4) I (4) (3 ) (4)

CHAPTER IV SUMMARY, DISCUSSION, AND CONCLUSIONS

Summary of the Study

This study was an investigation into the relationship

of Myers-Briggs Type Indicator preference scores to the age

of the respondents. Previous investigators have reported

that the MBTI preference scores are associated with age in a

manner that is not predicted by previous discussions of

psychological type and which appear to be inconsistent with

Jung's type theory. In this study I have attempted to find

a theoretical basis for associations of MBTI scores with age

and to correct methodological errors of the previous

analyses. Additionally, I sought a non-theoretical explan-

ation for the reported associations by examining the current

scoring procedures of the MBTI. I proposed that the

response weights employed for all respondents may not be

equally appropriate at all ages.

The Myers-Briggs Type Indicator (Myers & McCaulley,

1985) is a self-report questionnaire explicitly designed to

indicate the preferences of conscious orientation described

by Jung (1971) in his theory of psychological types.

Scoring the MBTI yields preference scores on four scales:

141

142

Extraversion-Introversion (EI) , Sensing-Intuition

Thinking-Feeling (TF), and Judgment-Perception (JP).

(SN) ,

Each preference score consists of a letter and a

number, e. g., E37, F15. The letter portion indicates the

direction of the preference and the number reflects the

strength or consistency with which the respondent has

reported the preference. Myers (1975) cautioned that the

interpretation of the number portion be limited to the

probability that the respondent has correctly reported the

preference and the relative importance of the preference to

the respondent, until further correlates aer empirically

found. Subsequently, Myers and other researchers have

investigated correlates of the MBTI scale scores with other

scores of abilities, aptitude, attitude, interests, values,

and traits of personality (see review in Myers & McCaulley,

1985) and have reported reliable, but generally moderate,

associations in support for the theoretical concurrent

validity of the scales.

Clearly, Myers did not intend for the MBTI scales to

be considered an exact measure of type preferences, but

rather a gauge for the preponderance of the evidence that

one function or attitude is preferred above its opposite.

However, the broader interpretation of the preference scores

is widely employed, with considerable empirical justifi­

cation, particularly in research analyses.

143

According to Jung's theory, consciousness is oriented

by either a function of perception (sensing [5] or intuition

[N]) or a function of judgment (thinking [T] or feeling [F])

that operates in one of two attitudes (extraversion [E] or

introversion [I]). In this study the term attitude-function

refers to the combination of those attitudes and functions,

such as extraverted sensing or introverted sensing. One

attitude-function, called the dominant function, is said to

dominate or lead the orientation of consciousness. A second

attitude-function, called the auxiliary function serves as a

complementary orientation. The auxil iary is complementary

with respect to attitude--if the dominant attitude-function

is extraverted, the auxiliary will be oriented toward

introversion; if the dominant attitude is introverted, the

auxiliary will be oriented toward extraversion. Further,

the dominant and auxiliary attitude-functions are

complementary with regard to judgment and perception. If

the dominant is a perceptive function (sensing or

intuition), the auxiliary must be a jUdging function

(thinking or feeling), and vice versa. For example,

extraverted sensing may be complemented by either

introverted thinking or introverted feel ing . The theory

assumes that the dominant attitude-function is the first to

develop; the auxiliary attitude-function develops later and to a lesser degree than does the dominant attitude-

function. The combination of the dominant and auxiliary

144

attitude-functions (such as, introverted thinking with

extraverted sensing) describes an individual's psychological

type.

Jung never presented an explicit developmental theory

and those presented by his followers did not deal with

psychological type. However, the central theme of his

theory, individuation, is defined as the psychological

development of adult life. Some prominent Jungian theorists

(Meier, 1971; Fordham, 1972) argue that the dynamic inter­

actions of the dominant and auxiliary attitude-functions

described in the type theory are the first steps of the

process of individuation. A central process of individ-

uation is the disengagement of consciousness from its iden­

tification with the ego-complex, which retains its asso­

ciation with the dominant and auxiliary attitude-function.

Therefore, the early importance of the dominant and

auxiliary attitude-functions is lessened as the individual

moves toward individuation in later life.

The individuation process, which Jung believed became

prominent at the mid-point of an individual's life (age 35

to 45), is a more extensive recapitulation of the process in

adolescence by which the auxiliary attitude-function within

the ego-complex emerges. Although, the dominant attitude­

function retains control, the auxiliary becomes increasingly

important in the individual's adaptation to the demands of

reality. According to Myers, such changes in the importance

145

of the attitude-function would be reflected in the MBTI

preference scores.

The theory assumes the changes of individuation occur

gradually over a period of time, and in some individuals

development may stop before the individuation process takes

hold. For the purposes of this study, I assumed that the

behavioral changes of the individuation process might be

reflected in the ways individuals emphasized or de­

emphasized their preferences when they responded to the

MBTI. The changes predicted, then, depended on the position

of the attitude-function as dominant or auxiliary at

different stages of the developmental cycle. The analyses

included eight separate age groups from 15-17 to GO-and­

over. In addition, development was considered to fall into

three main time spans: stage one included the ages 14 and

under (which did not enter into the analyses because of the

unreliability of the MBTI at younger ages and the

unavailability of data); stage two, the ages 15-39; and

stage three, the ages 40 and over. Preference scores

associated with the dominant function were expected to

increase to a small degree between adolescence and mid-life;

it was assumed that as the auxiliary developed, the hold of

the dominant function would be somewhat relaxed, and that

this relaxation would appear in a tendency of the preference

scores related to the dominant to remain constant or even

decline after mid-life. Preference scores associated with

146

the auxiliary were predicted to increase to a greater degree

from adolescence to mid-life (stage two) and to continue to

increase, but to a lesser degree, thereafter. Such changes

should appear clearly only when the distinction between

dominant and auxiliary attitude-functions can be maintained

in the analyses.

I found differences in mean MBTI preference scores of

respondents at various age levels that suggest that the

consistency of the preference, and possibly the direction of

the preference, may be associated with the age of the

respondent. Other investigators (Bloch, 1978; Schaeffer,

1974) have reported similar findings. still other

investigators (Gray, 1947b; Gray, 1948; Driver, 1974) have

reported age and degree-of-preference associations using

another independently developed Jungian type indicator, the

Jungian Type Survey (JTS) (Wheelwright, Wheelwright, &

Buehler, 1964). Across all these studies of type prefer­

ences and age, reliable positive associations have been

reported for introversion twice, for extraversion once, for

sensing twice, never for intuition, for thinking once, never

for feel ing , for judgment once, and never for perception.

These findings are not consistent with Jung's theory,

however, I proposed that methodological errors and failure

to take Jung' s theory fully into account may have caused

erroneous results.

147

The investigators who reported reliable associations

between MBTI scores and age did not select their samples to

control for the differences of the frequencies of the 16

possible types at different age levels. Thus, the reported

results could be an artifact of sampling bias. For example,

if more introverts at older ages than at younger ages were

included in the samples, the EI preference score would

necessarily increase. Furthermore, the previous investi­

gators did not distinguish preference scores associated with

dominant attitude-functions from those associated with

auxiliary attitude-functions. In theory, the consistency

with which an individual reports a preference at different

stages of life may be related to whether the preference is

dominant or auxiliary. New analyses were required to

correct the deficiencies of the previous studies.

One possible explanation for the reported associations

between age and MBTI preference scores might lie in sampling

bias by the investigators reporting the associations.

Another source of the associations might lie in the scoring

procedures of the MBTI. The initial development of the MBTI

mature adults were sampled to obtain the scale and item

statistics

revisions

that time

necessary to calibrate scales.

of the scales required larger samples,

were only available at younger ages.

Subsequent

which at

Items or

item responses that predicted type preferences more effec­

tively for mature adults may have been discarded in the

148

process of scale revisions, or may have been assigned

weights more appropriate for younger respondents. This

possibility could be investigated by replication of Myers's

item analyses separately for respondents at different age

levels.

I undertook this study to test theoretical predictions

of changes in consistency of preferences at different life

stages, and to investigate the possibility that persons of

different ages might respond differently to specific items

of the MBTI. I took advantage of a collection more than

250,000 MBTI records in the MBTI data bank maintained by the

Center for Applications of Psychological Type (CAPT). I

performed the analyses on a sample of 21,696 MBTI records

drawn randomly from the MBTI data bank. I selected the

sample to include equal numbers of each of the 16 MBTI types

for each gender at each of eight age groups (15-17, 18-20,

21-24, 25-29, 30-39, 40-49, 50-59, 60 and over). As may be

seen in Table 6 above, the frequencies of respondents at the

eight age levels differed widely; many more respondents at

the younger ages were available.

The research sample was used in a series of analyses:

analyses of variance by age groups and developmental stages,

regression analyses of trends of association of scores with

age, and item analyses controlled for age. The analyses

included a series of permutations of type combinations

related to the theoretical issues under examination.

149

Trends of association were predicted to occur between

specific MBTI preference scores and the respondents' ages

according to whether the preference score was associated

with the dominant or auxiliary attitude-function, and also

whether the respondents' ages fell into the range of

adolescence to mid-life (stage two: age 15 to 39 years) or

the range of mid-life and beyond (stage three: age 40 and

over). The trends were predicted to occur in patterns that

were related to psychological type theory by a simple model:

preference scores related to the dominant attitude-function

were expected to increase slightly from adolescence to mid­

life, and thereafter to remain constant or decline;

preference scores related to the auxiliary were expected to

increase moderately during the period from adolescence to

mid-life, and thereafter to increase only slightly. Because

of the large number of analyses I performed in this study, I

rejected the null hypotheses only when the probability of

the E ratio was .005 or less. The use of such a

conservative test decreased the risk of type I error but at

the same time increases the risk of type II error,

especially among the analyses for stage three, where the

estimates of the variances (being based on much smaller

samples) were less stable.

Extraversion declined with age among female extraverts

as predicted, but remained constant among males. For both

sexes, introversion, contrary to prediction, showed reliable

150

increases during stage two and remained constant thereafter.

These findings are more congruent with the interpretation of

a trend among both extraverts and introverts to increase

their reported introversion with increasing age than they

are congruent with my predictions.

Myers devised the MBTI JP scale to be a method for

identifying the dominant and auxiliary functions. She

derived it from Jung's theory, but it is not a direct

measure of Jung's constructs.

tion that preference scores

For this reason, the predic­

for judgment and perception

would increase with age have less theoretical support than

predictions for EI, SN, and TF.

Samples of respondents reporting a preference of the

judging attitude showed the predicted increases of the JP

preference score during stage two, and the two jUdging

samples predicted to show no JP score increase during the

third stage did not increase. However, two other jUdging

samples predicted to have an increase during the third stage

did not. Four samples reporting a preference for the

perceptive attitude were expected to show increases in

preference scores with age; contrary to predictions, three

of the four samples showed a reliable decrease in their JP

preference

perceptive

scores.

samples

preference scores.

During the third

were predicted to

stage, two of the

increase their JP

No reliable trends occurred. In

general, where trends occur between age and the JP

preference scores,

judging increases

151

it appears that

and the preference

attitude decreases with age.

the

for

preference for

the perceptive

Preference scores were expected to increase with age

in eight samples with either sensing or intuition as the

dominant function. six of the eight samples had reliable

increases in their SN preference scores during stage two; in

stage three no reliable trends emerged. However, the

relative size of some of the standardized regression

coefficients (BETA) among the stage three results suggests

the possibility of type II error.

As was predicted for samples reporting auxiliary

sensing or intuition, seven of eight samples had

significant increases in their SN preference scores in stage

two, and the expected positive trends did not occur during

stage three. Again, the relative size of some of the BETA

values suggests the possibility of type II error.

Of the samples of respondents reporting a preference

for dominant thinking or feeling, only two (male and female

extraverted thinking samples) had an increase of the TF

preference scores during stage two. No reliable trends

appeared during stage three. Some of the BETA values for

these analyses also suggest the possibility of type II

error.

For samples where thinking or feeling were auxiliary

functions, only two samples (male extraverted thinking and

152

male introverted thinking) had expected increases of the TF

preference scores during stage two.

trends did not emerge for stage three.

error may be suspected.

The expected positive

Once again, type II

In summary, significant support for the proposed model

appeared only among analyses of the SN preference scores.

Reliable associations found between the EI preference scores

and age, and between the JP preference scores and age, do

not fit expectations derived from theory. Rather, these

results are more consistent with the interpretation of

extraversion-introversion and judgment-perception as

continuously distributed attitudes that shift with

increasing age toward introversion and judgment, respec­

tively. This finding is not consistent with the underlying

theory of psychological types. Furthermore, because these

results were obtained from a carefully selected samples, the

lack of correspondence between expectations and the results

obtained does not appear to be accountable to biases in the

selection of MBTI records for the analyses.

The focus of the next series of analyses was on the

MBTI items themselves. The item analysis procedures that

Myers used to evaluate the predictive value for each of the

item responses are not associated with a statistical test to

estimate the probability of the results obtained. Myers

assigned weights according to the value of a prediction

ratio (PR). The PR was used to establish item weights of

153

zero, one, or two for each response. The replication of

Myers's procedures in this study generated prediction

ratios; these could not be matched directly with prediction

ratios obtained by Myers, because she did not publish the

ratios she obtained. Therefore, for my analyses I assigned

weights of zero, one, or two to the data bank samples and

compared these to the published weights. When the

calculated PR for a response was higher than the maximum for

its assigned weight, the current scoring key for that

response was considered under-weighted for its predictive

value. When the calculated PR for a response was lower

than the minimum for its assigned weight, the current

scoring key for that response was considered over-weighted

for its predictive value. In the analyses, therefore, if

more responses are under-weighted than are over-weighted for

a preference, the present scoring keys for that scale under­

estimate that preference. Conversely, if more responses are

over-weighted, the present scoring keys for that scale over­

estimate that preference.

The MBTI preference scores are calculated from the

difference between the points for responses related to one

preference and the points for responses related to the

opposite preference. Thus, overestimation of one preference

results also in underestimation of its opposite (the MBTI

scales are all ipsative). The result of the item analyses

most directly related to the problem of the study (trends of

154

association between the MBTI preference scores and age) is

the degree by which a preference may be overestimated at one

age (preference score too high) and underestimated at

another age (preference score too low).

The actual effects on the trends of association--

caused by the differential mis-estimation by age for the

consistency of the attitude-function preferences--could only

be determined by re-scoring the scales with response weights

revised according to the current calculations of the

prediction ratios, which is beyond the scope of this study.

However, the analyses indicate that the degree to which a

preference score would be affected is dependent on the age

of the respondent. After response weight revisions the

subsequent trends of association between the preference

scores and the respondents' ages would probably be changed.

Item analyses calculated for the total sample showed

that many responses did not fall into the range for their

assigned weight. The current response weights tend to over-

estimate the preferences

the judging attitude for

feeling among females.

for extraversion, intui tion, and

both sexes; and to overestimate

Because Myers's item analysis procedures do not have

an associated test statistic, I supplemented the analyses

with statistical tests of association between the

frequencies of the responses for each item and the eight age

groups. For each item I tabulated a contingency table of

155

responses by age groups for each sex. A two-tailed chi-

square test of association was applied to each table. I

rejected the null hypothesis (of no association) if the

probability of the chi-square statistic was .05 or less.

I rejected the null hypothesis for 19 of 22 items on

the EI scale for both genders, for 20 of 26 SN items for

females, for 20 of 26 SN items for males, for 17 of 23 TF

items for females, for 21 of 23 TF items for males, and for

21 of 24 JP items for both sexes. Thus, the choice of

response is associated with the age of the respondent for a

large majority of the items on each of the MBTI scales.

Discussion

This section is a discussion of some limitations of

the study, and of the relevance of the reported results for

understanding changes in the consistency of respondents'

reported preference for psychological attitude-functions

that are associated with the their age.

The data analyzed in this study were drawn from cross-

sectional samples. The central concern of the study is to

understand better how individuals may change the consistency

of their reported type preferences as they grow older. A

sample of longitudinal data would have made possible a more

rigorous test of the predictions, but such a sample does not

exist and cannot reasonably be obtained. Using a cross­

sectional sample implies an assumption that the mean

preference scores currently obtained from respondents in the

156

younger age groups are directly comparable to the mean

preference scores that would have been obtained from the

older respondents had the data been collected from them

when they were the age of the younger group. In making this

assumption I am ignoring differences of social climate

between the times when each group was of an equivalent age.

Uses of language, social attitudes, and other factors that

may alter the attractiveness of a particular response,

phrased in a particular way, may change considerably in a

generation or less.

guish the effects

Cross-sectional sampling cannot distin­

of social change from the effects of

maturation. In this study, those effects are confounded.

The limitation of using cross-sectional data may be an

inherent difficulty with the MBTI items, since Myers

intentionally selected items that reflect the ordinary

choices of everyday life, items that do not require the

respondent to take an extreme position. These may also be

the choices most easily influenced by the social climate and

by the choices of one's peers.

The effects being studied are small effects, although

the statistical analyses show them to be very reliable. The

evidence for their statistical reliability is demonstrable

because the large numbers of MBTI records studied allowed me

to obtain very stable estimates of the sample variances.

However, a great disparity among the frequencies of

respondents in the eight age groups occurred the sample.

157

One effect of this disparity was that data from analyses of

the younger respondents (who were more numerous) were far

more stable than that of the older respondents. This was

true both across the eight age groups, and wi thin both of

the developmental stages. A superior statistical design

might have been to limit the number of younger respondents

to equal the number of older respondents. However, such a

limitation would have reduced the stability of the variance

estimates for the younger age groups, making less detectable

the reliable trends that appeared. In most psychological

instruments careful study of these small effects would be of

little consequence, because a difference of a few points in

a respondent's scale position does not markedly change the

interpretation of the result. In the use of the MBTI as an

index of psychological type, fairly small differences in

scale position can shift the indicated type for respondents

whose scores lie near the mid-points. For this reason,

Myers took especial care correctly classify scores that were

near the mid-point. This study is a continuation of her

concern in that regard.

I performed a large number of independent statistical

analyses were performed on these data, and therefore adopted

an unusually small probability of the test statistics (.005)

for rejection of the null hypothesis. Some readers

inspecting the results, especially results for the third

developmental stage, may decide this probability level is

158

too conservative. For their consideration, the exact

probability of each result is provided with every test.

Although I designed the research to test specific

theoretical predictions, the findings that may have the most

practical consequences are the number of disparities between

the response weights used in the published MBTI scoring keys

and the predictive value of the responses indicated by the

analyses of the CAPT data bank records. Perhaps the need

for weight revisions of the TF scale in the mid-1970s should

have made this result more predictable. However, at that

time Myers examined all four scales and found sufficient

evidence for needed revisions only in the TF scale.

McCaulley (1985, personal communication) reported that

Myers did find discrepancies between prediction ratios and

assigned weights for the other scales, but concluded that

the results of her earlier analyses were more reliable

(having been based on better sampling procedures) .

It would be surprising if Myers's item analyses in

1977 had not found some discrepancies, considering the

number found in the current analyses. It is reassuring that

the drift found for the predictive values of the responses

for the EI, SN and JP has developed over 20 years rather

than 8 years. The smaller discrepancies detected for the TF

responses for males certainly reflect the more recent

corrections. The finding that the scoring weights now

underestimate thinking among females suggests that perhaps

159

the mid-1970s may not have been the best time to make

revisions in the weights of the female TF responses.

Because the data Myers analyzed at that time indicated that

thinking responses had become more popular among females,

she reduced their weight for some items. Those data may have

reflected too much of intense conflict concerning women's

changing roles. That is, women may now be more accepting

that the feeling function is neither simply a stereotyping

of traditional womens' roles, nor inconsistent with their

new roles; they may, therefore, be more readily endorsing

the feeling responses.

Table 41 shows the number of MBTI items per scale that

have one under-weighted or over-weighted response. For each

scale the number of items requiring revised weighting

indicates a clear need for a restandardization of the MBTI

scales. Some items now predict better than indicated by the

weights of the scoring keys, others, less well.

In general, the model developed in Chapter I, relating

changes of MBTI preference scores to the respondents' ages,

gained little support from the results reported here. The

SN preference scores most closely conformed to the

predictions based on the model. The model predicted that

preference scores associated with the dominant attitude­

function would increase reliably with age from adolescence

to mid-life and then remain constant. Also, the preference

scores associated with the auxiliary attitude-function would

160

increase more steeply than the dominant from adolescence to

mid-life, and show only slight increase thereafter. Even

with the SN scale, where the age differences most clearly

conformed to predictions, the preference score differences

in the strengths of the associations did not show the

auxiliary more closely associated with age than the

dominant; and for a number of samples the association of

the dominant was greater.

Analyses of the response weights did show noticeable

differences in the predictive value of numerous responses at

different ages. I observed both increases and decreases of

the sums of points that would be added to each scale by re­

weighting; other, non-linear patterns, appeared as well.

wi thout actually rescoring the entire data bank and re­

peating all the analyses, one cannot conclude that the age

differences among the response weight deviations account for

the pattern of associations between preference scores and

age found in the regression analyses.

ciations would probably be different.

The pattern of asso­

Speculating on the

basis of visual inspection alone, the trends of association

with age for sensing, intuition, thinking, feeling, and

perception are likely to increase, and the trends for

extraversion and introversion to decrease. Such a pattern

would not be incompatible with the type theory, although it

would not be consistent with the model I presented in this

study.

An inspection of

preference scores by age

161

the progression of the mean SN

showed distinct patterns related

to whether the preference was associated with an extraverted

or an introverted attitude for those samples. Differences

were also apparent between some samples depending on whether

the preference was dominant or auxiliary. If similar

patterns appear in future studies, their existence will

provide support for the theoretical distinction Jung made

between the extraverted and introverted attitudes and

between the dominant and auxiliary functions. Several of

the reliable associations reported in this research support

the possibility that further work will support Jung's

distinctions.

Conclusions

Sufficient numbers of MBTI item response weights on each of

the MBTI scales show deviation from their predictive

value to warrant a restandardization of all four scales.

Sufficient numbers of MBTI item responses differ in their

predictive value at different age levels to make

separate response weighting by age level desirable.

Using current scoring weights, reliable differences among

the means of the MBTI preference scores occur at

different age levels, and reliable trends of association

occur between the respondents' ages and preference

scores.

162

The differences of predictive value of the item responses at

different age levels influence the observed differences

among the means and the trends of association .

Revision of the MBTI item response weights would alter the

pattern of trends of association between MBTI preference

scores and the respondents' ages from those trends

reported in this study.

Recommendations For Further Investigation

In this research I attempted to test theoretical

predictions and to investigate the possibility of changes

in MBTI item weights. The changes in the predictive value

of the MBTI item responses that have occurred among all age

groups made it impossible, within the scope of this study,

to separate out these changes from the effects of the

developmental stages. Thus, further research is needed to

control separately for effects of both generational and age

differences.

The predictive value of the item responses does appear

to change with time. Therefore, the item analyses should be

replicated at intervals in order to determine when restan­

dardization is required.

Although, large samples of high school and college

students, and other young adults were available for analyses

in this investigation, the number of respondents aged 50

years and older cannot be considered sufficient for

normative studies. The use of the MBTI with older

163

populations is increasing rapidly. The findings from the

present study indicate the need for age-related normative

data for older samples. until such norms are available,

investigators and counselors interpreting the MBTI scores of

respondents over age 50 should give careful consideration to

the fact that the scales are presently normed on high school

students.

164

Table 44. Number of MBTI Items Having out-of-Range Response Weights

Scale Females Males Genders Combined Items on Scale

EI 15 14 14 22

SN 12 9 9 26

TF 11 3 NA 23

JP 13 12 13 24

APPENDIX RESULTS FROM THE PILOT STUDY

Table 42. Summary Anovas By Age Groups: Pilot Study

Females Males Preference F Ratio Probability F Ratio Probability

Extraversion 1. 929 .0608 33.168 .0000

Introversion 7.475 .0000 13.787 .0000

Sensing 63.700 .0000 57.214 .0000

Intuition 25.116 .0000 54.795 .0000

Thinking 89.603 .0000 24.946 .0000

Feeling 1.876 .0691 7.908 .0000

Judgment 110.121 .0000 72.264 .0000

Perception 7.600 .0000 9.685 .0000

165

35 34 33 32

P 31 R30 E29 F~

166

Extraversion­Introversion

Pilot Study N = 52,848

M E 26 E R 2S A E 24~---­N N 23

(22 E E 21

I S ~ ( 18 o 17 R 16 E 15

14 13 12

••••••• f • •••• :.:.~:.:.:.~_,.. ... _____ ••

~~~-~~------------ .... ---r~~·· . . . . .~"~ --.::~.....::::::...--.--=.-=-:.:..-==-::.. ':::'~' '-' - -- ..... .. _.:~~.r-:=~:;..:... ' ___ . ______ - ... _------_ ... -- ~ .....

11+------+------+------+------+------+------+-----~

1 2 3 4 5 6 7 8 Age Group

- Female --- Female .- Male ... Male Extraverts Introverts Extraverts Introverts

Figure 12. Mean EI Preference Scores: pilot Study

35 34 33 32

167

Sensing-Intuition Pilot Stuay N = 52.848

P 31 R 30 __ . E 29 ___ .--.

F~~ ~.--. M E 26 . /' E R 25 /-A E 24 b- ..... :. .. _ .. ~':"-:7~:::7~-:"h_4..a .. _ N N 23 _-.,=-", __ ~'7"' ---::::.;-:-. ' .. . .. ~ ~ - - - - ... 7' ~ :- -. -. -

~ ... ...... --~ .. -C 22 . __ ...• .._ .. -S E 21 .J....:...a. __ - ........

N S ~~ _~~:..:.~._-- .... -----c 18 o 17 R 16 E 15

14 13 12 ll+------+------+------+------+------+------+-----~

1 2

- Female Sensing

3 4 5 Age Group

-.- Female .- Mole Intuition Sensing

6

. .. Mole Intuition

Figure 13. Mean SN Preference Scores: pilot Study

8

35 34 33 32

P 31 R 30 E 29

28 F 27

168

Thinking-Feeling Pilot Study N = 52,848

--'~ M E 26 E R 25 A E 24 N N 23

~' '~ /' ',--._.-

Czz T E 21 ,..,-'

-,/.

F 20 5 19 C 18

/' ---~:::::-~~~---------------- ----------------------.. --.. --. --. -:--~--=---..-.£.:::..:..-_-........

o 17 R 16 .. . .. . . . ~ E 15--14 ...... -. . -. ~---.............. ' .. ' ............ ~ ......... ~ , .. , ..

13 12 11

1 2

- Fgmale Thinking

3

--- Fgmale FQeling

I I 4 5 6 7 ~gg Group

.- Male ... Male Thinking Fegling

Figure 14. Mean TF Preference Scores: Pilot Study

I 8

35 3. 33 32

P 31 R 30 E 29

F~ M E 26 E R 25 A E 2. N N 23

C 22 J E 21

P 5 ~ C 18 o 17 R 16 E 15

1. 13

169

Judging-Perceptive Pilot Study N = 52,848

--,-,~

---' ~' /' --,--' /'

,/ ,---

_____ '~ ~ ~ ~ ~,;.:.:..:..:..:..:..:..:..:..:..:.::- '..wr",-,-,-,- - ~ ~ ~ ~ ~~ ~ ~:-"":':':.:. :"_'_,_, __

12 11+------+------+------+------+------+------+------;

1 2

- Female Judging

3 4 5 AgQ Group

-_. Female Perceptive

0- Male Judging

6 7

.. , Male Perceptive

Figure 15. Mean JP Preference Scores: pilot Study

8

170

REFERENCES

Bloch, D. L. (1978). The developmental properties of Jungian psychological type as measured by the Myers­Brigg Type Indicator (Doctoral dissertation, Mississippi State University, 1978). Dissertation Abstracts International, ~, 1429A.

Driver, J. D. (1975). Personality differences in the elderly as a function of type and length of residence (Doctoral dissertation, University of Maryland, 1974). Dissertation Abstracts International, 2§ (03), 1860A -1861A.

Edinger, E. F. (1968). An outline of analytical psychology. Quadrant,~, 1-12.

Fordham, M. (1972). Note on psychological types. The Journal of Analytical Psychology, 17 (2), 111-115.

Gray, H. (1946). Jung's psychological types in relation to occupation, race, body-build. Stanford Medical Bulletin, i (3-4), 100-103.

Gray, H. (1947a). Jung's psychological types: Meaning and consistency of the questionnaire. The Journal of General Psychology, 37, 177-186.

Gray, H. (1947b). Psychological types and changes with age. Journal of Clinical Psychology, ~ (3), 273-277.

Gray, H. (1948). Jung's psychological types in men and women. Stanford Medical Bulletin, Q (1), 29-36.

Gray, H., & Wheelwright, J. B. (1945). Jung's psychological types, including the four functions. The Journal of General Psychology, ~, 265-284.

Gray, H., & Wheelwright, J. B. (1946). Jung's psychological types, their frequency of occurrence. The Journal of General Psychology, 34, 3-17.

Grenade, G. (1985). MBTI item analysis program [Computer program]. Gainesville, Florida: Center for Applications of Psychological Type, Inc.

Hartmann, H. (1958). Ego psychology and the problem of adaptation (D. Rapaport, Trans.). Journal of the American Psychoanalytic Association Monograph Series, 1. New York: International Universities Press. (original work published 1939).

171

Jacobi, J. (1973). The psychology of C. G. Jung (8th ed.). New Haven: Yale University Press.

Jung, C. G. (1961). Memories, dreams, and reflections (R. Winston, & C. Winston, Trans.; A. Jaffe, Ed.). New York: Vintage Books.

Jung, C. G. (1968). Analytical psychology: Its theory and practice (The Tavistock lectures). New York: Vintage Books. (original work published 1935).

Jung, C. G. (1971). Psychological types (H. G. Baynes, Trans.; R. F. C. Hull, Rev.). In H. Read, M. Fordham, & G. Adler (Eds.), The collected works of C. G. Jung, Bollingen series xx, § (whole volume). Princeton, N. J.: Princeton University Press. (Original work published 1921) .

Kainz, R. I., & McCaulley, M. H. (1980). Item omissions and MBTI split-half reliabilities]. Raw data analysis. Reported in Myers, I. B., & McCaulley, M. H. (1985). Manual: The Myers-Briggs Type Indicator (2nd ed.). Palo Alto, Calif.: Consulting Psychologists Press.

Kris, E. (1934). The psychology of caricature. Psychoanalytic Explorations in Art. New York: International Universities Press.

Lundberg, M. (1975). Developmental aspects of the Myers­Briggs Type Indicator. Paper presented at the first national conference on the Myers-Briggs Type Indicator, University of Florida, Gainesville, Fla., October, 1975. (Available from the Center for Applications of Psychological Type, 2720 N. W. 6th st., Gainesville, Fla., 32607.)

Marshall I. N. (1968). The four functions: A conceptual analysis. The Journal of Analytic Psychology, 13 (1), 1-32.

McCaulley, M. H. (1978). Applications of the Myers-Briggs Type Indicator to medicine and other health professions (Monograph I, Contract No. 231-76-0051, Health Resources Administration, DHEW). Gainesville, Florida: Center for Applications of Psychological Type.

Meier, C. A. (1971). Psychological types and individuation: A plea for a more scientific approach in Jungian psychology. In J. B. Wheelwright (Ed.), The Analytic Process (pp. 276-289). New York: G. P. Putnam's Sons.

172

Mosak, H. H., & Dreikurs R (1973). Adlerian psychotherapy. In Corsini, R., Current Psychotherapies (pp. 85-118). Itasca, Illinois: F. E. Peacock Publishers.

Myers, I. B. (1961). Construction of the type indicator: Form Zero to F. Unpublished manuscript.

Myers, I. B. (1975). Manual: The Myers-Briggs Type Indicator. Palo Alto, Calif.: Consulting Psychologists Press. (Original work published 1962).

Myers, I. B. (1977). Supplementary manual: the Myers­Briggs Type Indicator. Palo Alto, Calif.: Consulting Psychologists Press.

Myers, I. B., & McCaulley, M. H. (1985). Myers-Briggs Type Indicator (2nd ed.). Calif.: Consulting Psychologists Press.

Manual: The Palo Alto,

Neumann, E. (1954). The origins and history of consciousness (R. F. C. Hull, Trans). New York: Harper Torchbooks. (original work published 1949).

Norusis, M. J. (1984). SPSS/PC: Release 1.1 [Computer program]. Chicago: SPSS Inc.

Schaeffer, S. (1974). A study of sex and age differentials in typologies as derived from test scores of 200 adults on the Jungian Type Survey and the Myers­Briggs Type Indicator. Unpublished manuscript.

Stricker, L. J., & Ross, J. (1964). Some correlates of a Jungian personality inventory. Psychological Reports, 14, 623-643.

Wheelwright, J. B., Wheelwright, J. H., & Buehler, J. A. (1964). Jungian Type Survey, the Gray-Wheelwright Test, (16th revision> Manual. San Francisco: Society of Jungian Analysts of Northern California.

BIOGRAPHICAL SKETCH

Richard Ivor Kainz was born in 1945 in Cedar Rapids,

Iowa, where he lived until moving to Windermere, Florida, in

1957. In 1963 he graduated from Lakeview High School,

winter Garden, Florida. After attending New College in

Sarasota, Florida, and the University of Florida, he

received a Bachelor of Arts in psychology from the

University of Florida in 1973. He worked as a psychology

technician in the Typology Laboratory at the University of

Florida until beginning doctoral studies in the Department

of Clinical Psychology in 1974. He served as a Clinical

Fellow in Psychiatry at Harvard Medical School during 1978-

1979 and completed a clinical internship in psychology at

The Massachusetts General Hospital and Erich Lindemann

Community Mental Health Center in 1979. He has continued to

do research in psychological type as: graduate research

assistant, research associate, and director of research, at

the Typology Laboratory and its subsequent incarnation, the

Center for Applications of Psychological Type, Inc. (CAPT),

in Gainesville, Florida.

theory of psychological

Indicator at CAPT.

He continues to research Jung' s

types and the Myers-Briggs Type

173

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy.

~in Barger, airman Professor of Clinical

Psychology

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy.

ffiJ:jA:' /7u C' 4~ Mary . McCaulley ASSisar:t Professor of

Clinical Psychology

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy.

Davis, Jr. Profe sor of Clinical

Psychology

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy.

~rJf1:4Qp7'15 Professor of Psychology

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy.

This dissertation was submitted to the Graduate Faculty of the College of Health Related Professions and to the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy.

December, 1985 Dean, College of Health Related

Professions

Madelyn Lockhart Dean, Graduate School