15

Click here to load reader

Social-ecological assessment of environments: Toward a two-factor model

Embed Size (px)

Citation preview

Page 1: Social-ecological assessment of environments: Toward a two-factor model

American Journal of Community Psychology, Vol. 7, No. 5, 1979

Social-Ecological Assessment of Environments:

Toward a Two-Factor Model 1

Martin Kohn 2

William Alanson White Institute of Psychiatry, Psychoanalysis and Psychology, New York City

Abraham M, Jeger

New York Institute of Technology, OM Westbury, New York

Martin B. Koretzky

Veterans Administration Hospital, North,port, New York

A study was conducted to test the usefulness of the two-factor model in assess- ing social environments. Subfects were resMents and staff at a residential treat- ment center for emotionally disturbed and/or delinquent boys and girls. Rating instruments consisted of (a) two parallel versions of Moos' Community-Oriented Programs Environment Scale {COPES), each worded so as to be appropriate to the setting being measured, and (b ) global ratings of satisfaction with the environ- ment. Principal component factor analyses, carried out separately on the youths' COPES-School and the youths' COPES-Cottage, yieMed two orthogonal but similar factors in each environment. In a comparison with Moos' three-dimen- sional formulation of the social milieu, the two factors strongly resembled two of Moos' dimensions but were more independent, had greater validity, and dis-

1 The authors are indebted to the staff of Hawthorne-Cedar Knolls for their willingness to assist in this study and for giving generously of their time. We are especially grateful to Mr. Herbert Cohen, Principal of the Union Free School District #3; to the unit super- visors for freeing the necessary staff time; Dr. Sylvia Halitsky, School Psychologist; and Mr. Stanley Weinberg, Director of Clinical Services. We also wish to express our appre- ciation to Mrs. Rochelle Jeger who assisted in the data analysis and to Mrs. Vera Kohn for her thoughtful and conscientious work in editing the manuscript.

2All correspondence should be sent to Martin Kohn, William Alanson White Institute, 20 West 74 Street, New York, New York 10023.

481

0091-0562/79/1000-0481 $03.00/0 © 1979 Plenum Publishing Corporation

Page 2: Social-ecological assessment of environments: Toward a two-factor model

482 Kohn, ]eger, and Koretzky

criminated between the two environments at a higher level of significance. The instruments derived from the factor analyses provide a simple procedure for assessing a variety of treatment programs and populations.

As community psychology begins to develop a data base for large-scale social interventions, the area of environmental assessment is likely to emerge as a cen- tral focus of attention. As Cowen (1977) noted in his recent Division 27 Pre- sidential Address entitled "Baby-Steps Toward Primary Prevention," if we are ever to design health-promoting environments, we must first determine and assess the major dimensions of environments and how they are related to behavior. Thus, Cowen (1977) identified two key areas "as ideal candidates for primary prevention: the analysis and modification of social environments, and compe- tence building" (p. 6).

A promising and parsimonious approach to environment assessment is provided by the two-factor model of social environments. In the course of his work in both educational and industrial settings, Stern (1970) identified two factorially derived dimensions as characterizing environments. He called the first class of environmental "press" (Murray, 1938) anabolic, by which he meant situ- ational elements which are potentially conducive to growth and self-enhance- ment. He labeled the second catabolic, which referred to the characteristics de- signed to maintain order and stability in the environment and which Stern saw as antithetical to personal growth.

Kohn (1975; 1977) conceptualized the two-factor model of environments along slightly different lines. He postulated that the first dimension relates to opportunities for personal development and satisfaction as well as environment- al supports and means to achieve these satisfactions (comparable to Stern's anabolic dimension). The second dimension deals with norms, rules, and pro- cedures to guide social interactions and ensure orderly group functioning; regu- lations are not necessarily constraints on personal development and effective- ,hess however, since without organizational stability, the individual is too pre- occupied with self-preservation to seek self-fulfillment.

Kohn showed that similar situational dimensions have been found by He- witt and Jenkins (1946) and Roff (1949), who analyzed home environments, and Prescott and Jones (t967), who studied the social climate of day care centers.

Whereas the two-factor model approach is largely rooted in the fields of personality and developmental psychology, a popular approach to environmental assessment in the field of community psychology is reflected in the work of Moos and his colleagues (e.g., Moos, 1974b). They developed a series of "social climate" scales in the context of the emerging field of social ecology (see Moos & Insel, 1974) to assess nine different social environments. These include treatment en- vironments, total institutions, educational environments, and community settings.

Page 3: Social-ecological assessment of environments: Toward a two-factor model

Social-Ecological Assessment of Environments 483

The Moos scales were designed to tap three dimensions (to be called clus- ters in the present article) hypothesized to be common to all social environments: Relationship, Personal Development, and System Maintenance. While Moos placed his work against the background of previous investigators who had studied organizational climate (see Insel & Moos, 1974; Moos, 1974b), his tri- partite conceptualization of major clusters and the subscales within each cluster was, for the most part, rationally derived.

In order to make meaningful differentiations, however, the subscales and clusters should also be relatively independent. Taking the Community-Oriented Programs Environment Scale (COPES) as an example, although the 10 subscales were found to differentiate between various treatment programs (Moos, 1974b), the independence of the subscales is questionable. In looking at the 10 × 10 correlation matrix presented by Moos (1974a) for his normative sample, 15 of the 45 correlations between subscales were .30 or higher (for both the member and staff groups), a highly significant value due to the large n's (373 for member and 203 for staff). Moos (1974a) interpreted his findings to show that "the ten dimensions measure distinct [emphasis ours], albeit correlated characteristics of members and staff perceptions of program atmosphere" (p. 7). Close inspec- tion of the data, however, reveals that not only were the intercorrelations among subscales comprising one cluster fairly large, but the correlations between subscales of different clusters were also quite high. The data thus raise a question not only about the orthogonality of the 10 subscales, but also about the con- ceptual "identity" of the three broad clusters.

To date, the only study that challenges the Moos conceptualization em- pirically was conducted by Wilkinson (1973). Having factor analyzed a revised form of Moos' Social Climate Scale, Wilkinson found only one underlying di- mension. This was labeled the Value factor, reflecting an evaluative item dimen- sion, which was found to correlate higher with each subscale (for both staff and resident subjects) than did the subscales amongst themselves.

While Wilkinson's study seriously questions the dimensionality of the Moos scales, its influence seems to have been negligible considering the conti- nued widespread uses of the Moos scales in their current form. A possible reason for this might be that Wilkinson's approach was atheoretical, thus offering no alternative framework. In contrast, the present approach of measuring the en- vironment according to the two-factor model has its roots in previous theoretical and empirical formulations both about the social milieu and about personality functioning.

The purpose of the present study was to reexamine Moos' clusters from the point of view of the two-factor model. It was hypothesized that factor analyses of the Moos COPES, administered in two settings, would show that the

data fit our two-factor concept better than Moos' tripartite formulation. We also expected to demonstrate that the factors would compare favorably with Moos'

Page 4: Social-ecological assessment of environments: Toward a two-factor model

484 Kohn, Jeget, and Koretzky

clusters with respect to orthogonality of measures, validity of results, and ability to discriminate between environments.

METHOD

Setting

The present study was conducted at Hawthorne-Cedar Knolls, a residen- tial treatment center for emotionally disturbed and delinquent boys and girls. The facility, which is administered by the Jewish Board of Guardians (JBG), is located on a sprawling campus in suburban Hawthorne, New York, and can house approximately 220 youths at a time. The youngsters come from varied ethnic, religious, and socioeconomic backgrounds.

The youths are referred to the center by the New York City Family Court and the Department of Social Services as well as by private sources. The average length of stay is approximately 2 years.

Two major environments at Hawthorne-Cedar Knolls are the residential units and the school. Youths live in homelike cottages which can accommodate up to 18 members of the same sex and age level. Their main contact here is with child care workers who are responsible for attending to the youths' routine physical, emotional, and recreational needs.

The school is situated on another part of the campus and operates as a semiautonomous, self-contained school district. Classes are small, with a teacher- student ratio of approximately 1:5. The school is coeducational; students are assigned to classes on the basis of age and ability.

The youths are generally aware of distinct differences between the two environments. Not only are cottages and school separated physically, but they also inspire different psychological identifications. In the cottage, a sense of familylike identity is fostered; this is not carried over to the school, since cottage- mates attend different classes. On the other hand, the school encourages a strong identification with its long-standing tradition of innovative education.

Subjects

Three groups of subjects were employed in the study: the youths, who live in the cottages and attend the school; the child care staff, who supervise the youths in the cottage setting; and the teachers at the Hawthorne-Cedar Knolls School. Some background information on each group is presented below.

Youths. Of the population at the facility at the time of the study, 128 rated the cottage environment, and 192 rated the school environment (this

Page 5: Social-ecological assessment of environments: Toward a two-factor model

Social-Ecological Assessment of Environments 485

figure includes a number of day students who did not live on campus). 3 Average tenure at the facility at the time of the study was 1.8 years, with a range from

.2 to 7.3 years. Residents were between 9 and 18 years old, with a mean age of approximately 15; 80% of the youths were male. Ethnic background included White (69%), Black (18%), Puerto Rican (8%), and Mixed (5%). Approximate ly 50% of the youths came from broken families. With respect to educational level o f the parents, over two-thirds o f the parents had completed high school, and 25% of the fathers and 22% of the mothers had had one or more years of college.

Child Care Staff. Of the 67 child care workers at Hawthorne-Cedar Knolls, ratings were obtained from 56 o f whom 40 were male. Their mean length of employment at the facility was 3.3 years, with a range from .1 to 19 years.

Teachers. Of the 51 instructors at the school, ratings were made by 46 of whom 27 were male. On average, they had been working at the school for 8 years, with a range from 1 to 20 years.

Instruments

Two types of rating instruments were used. A modified version of the 40- item (short form) COPES developed by Moos (1974a; 1974b) was the chief en- vironmental assessment tool. 4 In order to permit a factor analysis, the original True -Fa l se response choice was converted to a 4-point format (Always true = 4, Usually true -- 3, Usually false = 2, Always false = 1). On the basis o f suggestions from the administrative staff, the wording o f the items was modified slightly to increase relevance to the Hawthorne-Cedar Knolls populat ion. One of the i tems was replaced by a paral lel i tem from the longer Moos version because it was more appropriate for the setting, s

Two parallel versions of the instrument were devised. One was designed to assess the cottage psychosocial climate and will be referred to as-"COPES- Cottage"; the other was designed to assess the school's psychosocial climate and will be referred to as "COPES-School."

The instrument has 10 subscales, each of which consists of four items. To reduce response set, half the items of each subscate were keyed positively, and the other half were keyed negatively. The following characteristics are tapped

3 The overall response rate was not as high as it could have been, largely due to the low num- ber of respondents from the youngest and most severely disturbed residents. Other reasons for not responding included illness, absence, and refusal to complete the form. Except for the low proportion of younger respondents, there was no systematic selectivity of sub- jects which might distort the results.

4 Moos (1974a) reported significant correlations between the parallel subscales of the short and long form. He has also provided extensive psychometric data.

s Copies of the instruments used in the study are available from the senior author.

Page 6: Social-ecological assessment of environments: Toward a two-factor model

486 Kohn, Jeger, and Koretzky

by the 10 subscales, grouped according to Moos' three broad clusters (the illustrations are drawn from the cottage version of the instrument):

Relationship Cluster

1. Involvement: the extent of day-to-day social involvement of the youths (e.g., "In my cottage kids really try hard").

2. Support: the extent to which the youths are perceived as receiving sup- port or assistance from other youths and staff (e.g., "In my cottage kids do not help each other").

3. Spontaneity: the extent to which the expression of opinions and feel- ings is encouraged (e.g., "In my cottage kids keep their feelings to themselves").

Personal Development Cluster

4. Autonomy: the extent to which independence is encouraged (e.g., "In my cottage kids are expected to take leadership").

5. Practical Orientation: the extent to which planning for the future is emphasized (e.g., "In my cottage kids have a chance to learn things which will be useful in working for a living").

6. Personal Problem Orientation: the extent to which examination of pro, blems with staff and youths is encouraged (e.g., "In my cottage personal pro- blems are openly talked about by staff and kids").

7. Anger and Aggression: the extent to which residents are free to argue with each other and staff (e.g., "In my cottage kids criticize or make jokes about the staff'').

System Maintenance Cluster

8. Order and Organization: the extent to which neatness and activity planning are emphasized (e.g., "The cottage looks messy").

9. Program Clarity: the extent to which rules, procedures, and goal ex- pectations are clearly perceived (e.g., "The rules of my cottage are not clear to the kids").

10. Staff Control: the extent to which staff is perceived as determining the rules to control the youths (e.g., "When a schedule is made for a kid in my cot- tage, he must follow it").

In addition to the COPES, global assessments of the environments were obtained by asking the youths to rate their overall satisfaction with the cottage and with the school on a 4-point scale whose anchor points were Extremely satisfied = 4, Satisfied = 3, Dissatisfied = 2, and Extremely dissatisfied = 1.

Page 7: Social-ecological assessment of environments: Toward a two-factor model

Social-Ecological Assessment of Environments 487

Procedures

The youths completed the COPES-School form and the global rating of the school during a classroom period, Teachers provided group instructions, ex- plaining that the survey was part of a Jewish Board of Guardians project seeking consumer feedback on their facilities. A week later, the teachers completed the identical COPES-School form.

After 2 weeks, during the same class period, the youths filled out the paral- lel COPES-Cottage instrument and the global rating of satisfaction with the cottage. At the same time, the child care staff were provided with the identical COPES-Cottage form to give their perceptions of the cottage environment.

Administration time on each rating occasion was approximately 30 minutes, with the younger children taking somewhat longer to make the ratings.

RESULTS

Psychometric Characteristics of the Instruments

Factor Analysis. Principal component factor analyses were carried out separately on the youths' COPES-School and the youths' COPES-Cottage. Factors were rotated by means of the varimax method to yield maximum ortho- gonality. Seven different factor solutions were tried encompassing two- through eight-factor solutions. No clear breaking points in percentage of variance account- ed for were found. Since the first two factors seemed clearly to be in line with our theoretical formulation, and since additional factors accounted for little ad- ditional percentages of the variance, a two-factor solution was chosen, based on the two-factor rotation. The two-factor solution accounted for 28% of the communal variance of the COPES-School responses and 37% of the communal variance of the COPES-Cottage responses.

In order to obtain factor scores of maximum reliability and stability and also to achieve maximum discrimination between the two factors, only items which, within a setting, had (a) factor loadings above .30, and (b) low factor loadings on the second factor were retained for subsequent analysis. The items and their loadings on each factor are presented in Table I for the JBG School ratings and in Table II for the JBG Cottage ratings. The Moos subscale and cluster to which the items belong are also shown.

It is apparent that the same or similar items make up Factor I in both settings, and the same or similar items make up Factor II in both settings; the corresponding factors from the two environments were, therefore, given the same labels - namely, Factor I: Support-involvement vs. disinterest; Factor II: Order-organization vs. disorder-disorganization.

Page 8: Social-ecological assessment of environments: Toward a two-factor model

488 Kohn, Jeger, and Koretzky

Table I. Youth Factor Items on JBG School Instrument

Factor loadings Items Moos

(Moos subscale in parentheses) cluster a I b II c

Part A: Factor items

Positive items In this school students really try hard

to learn (Involvement) A .45 .01 In this school students who are more

together help take care of those who are less together (Support) A .37 .06

People here don' t hide things. They talk about what's bothering them (Per- sonal Problem Orientation) B .58 - .08

In this school students tell the staff what's o n their minds (Spontanei ty) A .32 .07

Personal problems are openly talked about by the staff and kids in this school (Personal Problem Orienta- tion) B .42 - .27

Students are proud of their school (Involvement) A .50 .03

The staff in this school want to know what happens to the students after they leave Hawthorne (Support) A .43 .03

Negative items In this school people keep their feel-

ings to themselves (Spontaneity) A - .30 - .10 This school is a dull place (Involve-

ment) A - .31 - .15 In this school students criticize or

make jokes about the staff (Anger/ Aggression) B - .38 .16

There is no group spirit in this school (Involvement) A - .55 .08

Part B: Factor II items

Positive item The staff in this school make sure

that the place is always kept neat (Order/Organization) C

Negative items The staff in this school have little

time to encourage the students (Support) A - .20

In school students can always leave the classroom without saying where they are going (Autonomy) B .11

The whole school program is unorganized (Order/Organization) C - .23

.22 .52

- .42

- .46

- . 4 4

Page 9: Social-ecological assessment of environments: Toward a two-factor model

Social-Ecological Assessment of Environments

Table I. Continued

489

Items (Moos subscale in parentheses)

Moos cluster a

Factor loadings

I b IIe

Part B: Factor II items

In school students are not punished by having their privileges taken away (Program Clarity) C .26 -.41

In this school students are expected to take leadership (Autonomy) B .03 -.31

In this school it's hard to find out what the program is about (Program Clarity) C -.20 -.46

In this school students who break the rules are not punished for it (Staff Control) C .01 -.43

There is no talk in school about what students will be doing when they leave Hawthorne (Practical Orienta- tion) B -.15 -.35

The school looks messy (Order/Organiza- tion) C -.29 -.42

The rules of this school are not clear to the students (Staff Control) C -.02 -.37

aMoos cluster: A = relationship; B = personal development; C = system maintenance.

bFactor I = support-involvement vs. disinterest. CFactor II = order-organization vs. disorder-disorganization.

It is also evident that the JBG (i.e., our) factor-analytically derived dimen-

sions of the environment are somewhat comparable to two of Moos' clusters,

Relationship and System Maintenance, especially in the school. In the JBG

School factor analysis, 8 of 1 t Factor I items come from the three subscales in

Moos' Relationship cluster, and 7 of 11 Factor tI items come from subscales in

Moos' System Maintenance cluster (see Table I). In the JBG Cottage factor ana-

lysis, the pattern is somewhat less clear, with 5 of 12 Factor I items belonging to subscales in Moos' Relationship cluster, and 4 of 9 Factor II items belonging to Moos' System Maintenance subscales (see Table II). Items from Moos' Per-

sonal Development subscales occur as often as items from the Relationship sub- scales on Factor I and as often as System Maintenance subscales on Factor II.

R e l i a b i l i t y . Factor scores were then calculated, with each item receiving equal weight. Scores on each factor based on the youths ' JBG School factor an- alysis were computed separately for each youth and each teacher; scores on each factor based on the youths ' JBG Cottage factor analysis were computed for

Page 10: Social-ecological assessment of environments: Toward a two-factor model

490 Kohn, Jeger, and Koretzky

Table II. Youth Factor Items on JBG Cottage Instrument

Factor loadings Items Moos

(Moos subscale in parentheses) cluster a I b II c

Part A: Factor I items

Positive items In my cottage kids really try hard

(Involvement) A In my cottage kids who are more together

help take care of those who are less together (Support) A

In my cottage kids have a chance to learn things which will be useful in working for a living (Practical Orientation) B

In my cottage kids' activities and routines are carefully planned (Order/Organiza- tion) C

In my cottage personal problems are open- ly talked about by staff and kids (Per- sonal Problem Orientation) B

Kids are proud of their cottage (Involve- ment) A

Kids in my cottage are expected to make specific plans for their future (Practical Orientation) B

The staff in my cottage want to know what happens to kids after they leave Hawthorne (Support) A

Kids in my cottage are made to feel they should not criticize the staff (Autonomy) B

Kids in my cottage are expected to discuss their personal problems with each other (Personal Problem Orientation) B

In my cottage if a kid fights with another kid he will get into trouble with the staff (Staff Control) C

Negative item My cottage is a dull place (Involvement) A

Part B : Factor II items

.45 - .08

.47 - .16

,57 - .11

,57 - .16

,49 - .02

.68 - .01

.58 .10

.33 - .10

.34 .15

.50 .07

.30 - .09

- .44 .06

Positive item The staff in my cottage make sure that

the place is always kept neat (Order/ Organization)

Negative items In my cottage kids can always go out

without saying where they are going (Autonomy)

In my cottage kids do not help each other (Support)

C .11 .65

B .13 -.45

A .18 - .42

Page 11: Social-ecological assessment of environments: Toward a two-factor model

Social-Ecological Assessment of Environments

,Table II. Continued

491

Items (Moos subscale in parentheses)

Moos clustela

Factor loadings

Ib IIC

Part B: Factor II items

In my cottage the staff do not ask kids about their personal problems (Personal Problem Orientation) B .13 -.37

Kids who break the rules in my cottage are not punished for it (Staff Control) C -.01 -.45

There is no talk in my cottage about what students will be doing when they leave Hawthorne (Practical Orientation) B .03 -.36

Staff members of my cottage argue with each other in front of the kids (Anger/ Aggression) B -.07 -.44

The cottage looks messy (Order/Organiza- tion) C .13 -.36

The rules of my cottage are not clear to the kids (Program Clarity) C -.12 -.54

aMoos cluster: A = relationship; B = personal development; C = system maintenance.

bFaetor I = Support-involvement vs. disinterest. c Factor II = Order-organization vs. disorder-disorganization.

each youth and each child care worker. 6 Split-half reliabilities were calculated for each of the factor scores. Reliability coefficients (Spearman-Brown corrected)

ranged from .56 to .70 on the school instrument and from .57 to .76 on the cottage instrument. These r values were deemed sufficiently high for our purposes.

Orthogonal i t y . To determine the extent to which the two factors were

independent of each other, Pearson product moment correlations were obtained

between the youths ' scores on Factors I and II (since factor analyses were based

on their responses). The correlation in both environments was .14, low enough to indicate mutual independence.

Next, scores based on Moos' Relationship and System Maintenance clusters

(to be called Cluster A and Cluster C, respectively) were computed, and these

scores were correlated. The r values for the Moos clusters were .34 in the school

and .45 in the cottage setting (both p < .001), indicating that the JBG factors

were more independent than the Moos clusters. The differences between the JBG and Moos correlations were significant (t = 2.04, p < .05 for the school environment, and t = 2.42, p < .01 for the cottage environment).

We did not carry out factor analyses of the responses of the child care workers and the teachers, but applied the factor-analyticaUy derived solutions of the youths' ratings to the staff.

Page 12: Social-ecological assessment of environments: Toward a two-factor model

492 Kohn, Jeger, and Koretzky

Satisfaction and Perception of the Environment

As a test of validity, the relationship between the youths ' global ratings of satisfaction with the school and cottage environments and their perceptions of the psychosocial climate in each environment was examined. Data are presented in Table III, bo th for the JBG factors and the Moos clusters.

Comparing Factors I and II, it may be seen that in bo th settings, Factor I was more highly related to satisfaction than Factor II. Only Factor I, however, reached significance in the school environment, whereas bo th factors were signi- f icantly correlated with satisfaction in the cottage. The data clearly indicate that satisfaction with an environment is more strongly related to the support and involvement available than to the sheer orderliness of the setting.

Correlations between satisfaction and the JBG factors were consistently higher than correlations between satisfaction and the Moos clusters, although none o f the differences between the two sets of correlations were statistically significant.

Differences Between Environments

To determine the abili ty o f the instruments to discriminate between the two distinct environments, statistical tests of significance were carried out for youth and staff perceptions separately. The results are summarized in Table IV.

Table III. Relationship Between Youth Satisfaction and Perception of Social Ch-

mate a

JBG Moos factors b clusters c

Satisfaction I II A C

School (n = 107) .23 d .18 .13 .10

Cottage (n = 126) .39f .29 e .39f .26 e

aHigh score = more satisfied and social climate perceived as better.

bJBG Factor: I = support-involvement vs. disinterest; II = order-organization vs. dis- o~der -disorganization.

CMoos cluster: A = relationship; C = system maintenance.

dp < .05. ep <<..01. f p <<..001.

Page 13: Social-ecological assessment of environments: Toward a two-factor model

Social-Ecological Assessment of Environments

Table IV. Differences Between Environments: Comparison of JBG Factors and Moos Clusters a

493

School Cottage Environmental F (repeated dimensions b X SD X SD measures)

Youth

(n = 192) (n = 128)

JBGI -7.76 4.01 -23.58 5 . 1 1 p<.001 JBG II 22.26 4.06 26.33 4.54 p ~< .001 MoosA .38 1.38 .04 1 . 5 0 p~<.02 Moos C 1.87 1.49 1.57 1.67 p < .02

Staff

t (indepen- (n = 46) (n = 56) dent groups)

JBG I -3.13 2.42 -19.05 3.70 p ~ .001 JBG II 24.17 2.90 31.65 3.35 p < .001 Moos A 1.89 .80 1.97 1.24 ns MoosC 2.38 1.20 3.28 1 . 4 4 p~<.01

a High score = environment perceived as better. JBG factor scores were standardized to correct for differing number of items.

bJBG I = support-involvement vs. disinterest, JBG II = order- organization vs. disorder-disorganization; Moos A = relationship, Moos C = system maintenance.

The JBG Factors I and II successfully differentiated between the two

settings for both youth and staff (p ~< .001) for all comparisons. The data show

a large difference between the school and cottage on Factor I, with the school perceived by both groups as offering more opportunities for personal growth and development than the cottage. On Factor II, the differences between the

environments were significant but not substantial for both groups.

The two Moos clusters discriminated between the settings for youth, but only Cluster C did so for staff. Both the absolute differences and the levels of significance were greater for the JBG factors (particularly Factor I) than for the Moos clusters.

DISCUSSION

The emergence of two factors from each of two separate factor analyses confirmed our hypothesis that the two-factor model provides a theoretically

meaningful and parsimonious framework for measuring social environments. Each of the two factor dimensions in the two settings could be assessed with a

satisfactory degree of reliability. Our results support the ideas of Kolm (1975;

Page 14: Social-ecological assessment of environments: Toward a two-factor model

494 • Kohn, Jeger, and Koretzky

1977) and the previous work of Stern (1970) who, using factor analytic tech- niques, obtained dimensions similar to ours in other environments. At the same time, the findings seriously challenge the unidimensional results of Wilkinson (1973).

Relative to Moos' Cluster A and C, the JBG factors were more orthogonal, provided a finer discrimination between the two environments studied (school and cottage), and showed at least as strong, or stronger, a correlation with satis- faction. The JBG scales are shorter and thus easier to administer and score, and they are also easier to interpret (using two factors in lieu of 10 subscales). The instruments provide a simple procedure for assessing a variety of programs, such as residential treatment centers, shelter and detention facilities, day treatment programs, and with diverse populations, sucla as the aged in day care centers or psychiatric patients in residential settings.

Although there was some similarity between our Factors I and II and Moos' Cluster A and C, respectively, numerous shiftings of Moos' items were re- quired before the two JBG factors emerged as they did. Moos' middle dimen- sion (Personal Development) did not hold up empirically.

Curiously enough, although Moos was clearly aware of Stern's (1970) empirical formulations (e.g., Insel & Moos, 1974), he forced Stern's subscales into his three-dimensional conceptualization. It may turn out that while the proposed two factors occur consistently across all environments, Moos' middle dimension may vary from setting to setting.

On a more general level, our study further illustrates the importance of assessing environments in community psychology practice. Perhaps this is most apparent from the pronounce d differences found between the school and cot- tage environments on Factor I. In informal discussions, the Hawthorne-Cedar Knolls administrators had conveyed the impression that the school was "better." The view was substantiated by the youths' global ratings of satisfaction, which were close to 3 on average for the school and close to 2 on average for the cot- tage (on a 4-point scale ranging from Extremely satisfied = 4 to Extremely dis- satisfied = 1). The factor items give a clue to the cause of satisfaction with the school: This residential treatment center fosters involvement and provides op- portunities for meaningful interpersonal relations to a greater degree in the school than in the cottage.

Two suggestions for future research follow from the current findings. First, more items designed to tap the two dimensions conceptualized by us should be written and a new factor analysis conducted. By including items which measure both the positive and negative poles of each factor (unequally repre- sented in the current version), the total percentage of variance accounted for would probably increase over those found in the present study. Since we were interested in examining the dimensionality of Moos' instrument in its current form, separate items were not written in the present study.

Page 15: Social-ecological assessment of environments: Toward a two-factor model

Social-Ecological Assessment of Environments 495

Second, there should be explorat ion o f the impact o f the two environ-

mental factors on personali ty dimensions and the ways in which the social situa-

t ion determines behavior (see Kohn, 1977). This research strategy would inte-

grate the two areas (social competence and environmental assessment) that Cowen (1977) targeted for "baby-steps toward primary prevention."

REFERENCES

Cowen, E. L. Baby-steps toward primary prevention. American Journal of Community Psy- chology, 1977, 5, 1-22.

Hewitt, L. E., & Jenkins, R. L. Fundamental patterns of maladjustment: The dynamics of their origin. Springfield, Ill.: D. H. Green, 1946.

Insel, P. M., & Moos, R. H. Psychological environments: Expanding the scope of human ecology. American Psychologist, 1974, 29, 179-188.

Kohn, M. The social systems meaning of the two-factor model of social-emotional compe- tence. Paper presented at the meeting of the American Educational Research Asso- ciation, Washington, E.C., April 1975. (ERIC Document Reproduction Service No. 127 400)

Kohn, M. Social competence, symptoms, and underachievement in childhood: A longitudinal perspective. Washington, D. C., Winston-Wiley, 1977.

Moos, R. H. Community-oriented programs environment scale manual Palo Alto, Calif.: Consulting Psychologists Press, 1974. (a)

Moos, R. H. Evaluating treatment environments: A social ecological approach. New York: Wiley, 1974. (b)

Moos, R. H., & Insel, P. M. lssues in social ecology: Human milieus. Palo Alto, CaliL: National Press Books, 1974.

Murray, H. A. Explorations in personality. New York: Oxford University Press, 1938. Prescott, E., & Jones, E. Group day care as a child.rearing environment. Unpublished

manuscript, Pacific Oaks College, 1967. Roff, M. A. A factorial study of the Fels Parent Behavior Scales. Child Development,

1949, 20, 29-45. Stern, G. G. People in context. New York: Wiley, 1970. Wilkinson, L. An assessment of the dimensionality of Moos social climate scale. American

Journal of CommunityPsychology, 1973, 1, 342-350.