Psychological /2000, Vol 12, No. I, 77-88

Copyright 2000 by the American Psychological Association, Inc.1040-3590/00/$5.00 DOI: 10.1037//1040-3590.12.1.77

Development and Validation of the Stephenson MultigroupAcculturation Scale (SMAS)

Margaret StephensonUniversity of Massachusetts at Amherst

This article describes the development and psychometric evaluation of the Stephenson Multigroup

Acculturation Scale (SMAS). Three studies were conducted to describe its development and refinement,

examine its psychometric properties with 436 participants from 5 ethnic groups, and examine the

robustness of the factor structure with a new sample. Exploratory factor analyses generated a 2-factor

solution: ethnic society immersion and dominant society immersion. Item refinement resulted in a

32-item version of the SMAS. Findings indicated a robust factor structure across groups. Confirmatory

factor analysis indicated that the 2-factnr model provided a close approximation to the observed data.

Studies indicated high reliability and validity indexes. Findings support the role of acculturation as a

mediator between ethnic group affiliation and standardized assessment results.

As U.S. demographics continue to change, acculturation is in-

creasingly being viewed as an important variable that provides a

framework for understanding between- and within-ethnic group

differences. The importance of acculturation has not been evi-

denced in the psychological literature at large for a number of

reasons. First, the complexity of the construct of acculturation has

historically not been well understood. Second, this construct has

been applied only to selected ethnic minority groups, suggesting

that the process of acculturation applies to some groups and not to

others. Third, research across racial and ethnic groups in the

United States has historically focused on defining difference as

deficit. Thus, the focus of much research has been on the origins

of deficits rather than on questions pertaining to the roles of

experiences, environment, and culture in distinguishing pathology

from distinctiveness (Watts, 1994).

Conceptualization of Acculturation

The theoretical conceptualization of acculturation has shifted

from a simplified bipolar model (unacculturated to acculturated or

assimilated) to the recognition that acculturation is a complex,

multidimensional process of learning that occurs when individuals

and groups come into continuous contact with different societies.

This process appears to be common across many kinds of cultural

groups (immigrants, migrants, indigenous people, sojoumers, and

refugees), and voluntary as well as involuntary groups (Berry &

Sam, 1997). What varies is the course, level of difficulty, and to

some extent, the outcome of this process (Berry, 1996). Involve-

ment in one society does not necessitate a decrease in involvement

in another; therefore, individuals can assume a number of accul-

turation positions (Berry, 1980). On the basis of this theoretical

Correspondence concerning this article should be addressed to Margaret

Stephenson, Psychology Department, Tobin Hall, University of Massachu-

setts at Amherst, Amherst, Massachusetts 01003-7710. Electronic mail

may be sent to msl@psych.umass.edu.

conceptualization, assessment of acculturation entails measure-

ment of identification with the society of origin as well as with the

dominant society.

A widely accepted and cited framework proposed by Berry and

his colleagues conceptualized individual-level acculturation as a

multidimensional process of change that occurs when individuals

of differing cultural groups come into continuous contact (Berry,

1980; Berry, 1992; Berry, 1996; Berry & Kim, 1988; Berry &

Sam, 1997). These changes result in a variety of acculturation

positions determined according to how individuals deal with two

central issues. The first issue addresses retention of, or immersion

in, an ethnic society other than the dominant society. The second

addresses adoption of, or immersion in, the dominant society. The

negotiation of these two central issues results in four distinct

acculturation positions or modes of acculturation: assimilation,

integration, separation, or marginalization. Assimilation entails

moving away from one's ethnic society and immersing fully hi the

dominant society. Integration entails immersion in both ethnic and

dominant societies. Separation entails withdrawal from dominant

society and complete immersion in ethnic society, a process that

may be self-imposed or societally imposed (e.g., segregation).

Marginalization entails a lack of meaningful immersion in either

ethnic or dominant societies. Viewed from this framework, an

independent assessment of degree of immersion in dominant and

ethnic societies is needed.

Importance of Acculturation

The importance of measuring acculturation lies first in the

ability to provide an index of the degree of confidence that can be

assumed in interpreting standard assessment procedures. Measure-

ment of acculturation also serves as a tool to delineate the relative

contributions of dominant and ethnic society experiences in ob-

served differences between groups in research, assessment, and

clinical presentation (Dana, 1993). One can assume that many

traditional psychological theories are particularistic rather than

universal and that most traditional instruments were normed on

dominant group participants, reflecting a particular world view as

77

This

doc

umen

t is c

opyr

ight

ed b

y th

e A

mer

ican

Psy

chol

ogic

al A

ssoc

iatio

n or

one

of i

ts a

llied

pub

lishe

rs.

This

arti

cle

is in

tend

ed so

lely

for t

he p

erso

nal u

se o

f the

indi

vidu

al u

ser a

nd is

not

to b

e di

ssem

inat

ed b

road

ly.

78 STEPHENSON

well as a particular social context (Trickett, Watts, & Birman,

1994). It stands to reason that those individuals who more closely

fit particular demographics will perform differently than those who

deviate from them.

Numerous research findings have supported the assumption that

acculturation level impacts assessment results. For example, sig-

nificant differences have been noted on the Minnesota Multiphasic

Personality Inventory (MMPI; Hathaway & McKinley, 1943) pro-

files of ethnic minority group members compared with members of

ethnic majority groups (Dana, 1993; Dana, 1995; Montgomery &

Orozco, 1985). Montgomery, Arnold, and Orozco (1990) reported

that, when acculturation level was statistically controlled, differ-

ences in MMPI scores between ethnic minority and ethnic majority

individuals were reduced. This finding indicates the importance of

acculturation level in the interpretation of standardized tests.

Greene (1987) has emphasized that conclusions about real differ-

ences between groups must be carefully considered if relevant

variables are not controlled and analyzed. He also stated that

moderator variables are more important determinants of perfor-

mance than ethnic group status. It appears that assessing the role of

moderator-mediator variables in observed differences between

groups and clarifying an individual's identification widi a partic-

ular ethnic group are imperative (Greene, 1987; Zalewski &

Greene, 1996).

Similarly, numerous researchers have begun to address the

impact of acculturation on the concept of illness, expression of

symptoms, mental health functioning, entry into the mental health

system, and the therapeutic process (Atkinson, Casas, & Abreu,

1992; Atkinson, Thompson, & Grant, 1993; Matin, Gamba, &

Marin, 1992; Montgomery, 1992). These researchers have collec-

tively pointed to the importance of assessing the strength of

identification with a particular ethnic group in the recognition and

interpretation of clinical symptoms (Dinges & Cherry, 1995;

Grieger & Ponterotto, 1995; Keitel, Kopola, & Adamson, 1995).

Moreland (1996) suggested that individuals who are immersed in

dominant American society can be safely assessed with standard

interpretive procedures, whereas individuals who are not immersed

in dominant American society require alternative assessment pro-

cedures that are more culturally sensitive.

Second, acculturation is an important construct for making sense

of differences within ethnic groups. In the United States, the

process of acculturation has been applied exclusively to racial and

ethnic minorities and is assumed to be relevant only to those

groups. The assumption is that Whites and ethnic majority indi-

viduals are assimilated, whereas non-Whites and ethnic minority

individuals are unacculturated. Keeping in mind that the attitudes

of the dominant society toward particular groups will determine, in

part, the acculturation experience, process, and ultimate adaptation

of those groups (Berry, 1980), acculturation is likely not a process

relevant only to racial and ethnic minority individuals. In each

ethnic group there exists a wide array of within-group differences

that result from differing experiences in dominant American soci-

ety. The fact that an individual appears to be of a particular ethnic

or racial group does not mean that the individual is immersed in

one society rather than another (Dana, 1993). For example, a racial

minority individual may be more immersed in dominant American

society than a first-generation White European immigrant. It

would seem that all individuals, regardless of race or ethnic group

affiliation, should undergo some process of change in order to

adapt to a society different from their society of origin. Research

indicates that the process of adaptation is common across many

cultural groups (Berry & Sam, 1997). It would stand to reason that

testing a first-generation immigrant would be very different than

testing a third- or fourth-generation individual (Suzuki, Vraniak, &

Kugler, 1996), regardless of race or ethnic group affiliation.

Despite the documented importance of acculturation, research

studies have not examined common aspects of this process across

ethnic groups. Furthermore, the gap in the available assessment

instruments has limited research and has made widespread assess-

ment of this important variable impossible. There are currently

available a number of acculturation instruments designed for use

with specific Asian American groups (Matsumoto, Meredith, &

Masuda, 1970; Sodowsky & Carey, 1988; Suinn, Rickard-

Figueroa, Lew, & Vigil, 1987; Yao, 1979), specific Hispanic

American groups (Cuellar, Arnold, & Maldonado, 1995; Marin &

Gamba, 1996; Marin, Sabogal, VanOss Marin, Otero-Sabogal, &

Perez-Stable, 1987; Mendoza, 1989; Olmedo, Martinez, & Mar-

tinez, 1978; Padilla, 1980; Szapocznik & Kurtines, 1993), and

African Americans (Landrine & Klonoff, 1996). However, most of

these instruments do not measure acculturation as bidimensional,

and none are recommended for use with other groups. Some

researchers suggest that specific groups require individualized

attention hi order to attend to the particular cultural elements

specific to those groups (Rogler, Cortes, & Malgady, 1991). It is

indisputable that diverse groups will have differing experiences

rooted in their respective cultures; however, it is also likely that

there will be common experiences across acculturating groups.

Marin (1992) suggested that the acculturation process may be

perceived as occurring on three levels: the superficial, the inter-

mediate, and the significant. The superficial level involves, for

example, the learning and forgetting of historical facts and tradi-

tions, and changing diet to include foods from the dominant

society. The intermediate includes more central behaviors such as

language use and preference, degree of interaction within ethnic

and dominant societies, and environmental preferences such as

media. The significant level involves beliefs, values, and norms.

As Rogler et al. have noted, current assessment of acculturation

has not tapped this process on the significant level.

Although behavioral and attitudinal indices of acculturation are

not proxy variables for norms, values, and beliefs, Marin et al.

(1987) found a significant association between a behavioral index

of acculturation and familism (family collectivism or interdepen-

dence) in Hispanics. Research with the MMPI (Montgomery et al.,

1990), the Halstead-Reitan Neurological Battery (Arnold, Mont-

gomery, Castaneda, & Longoria, 1994), and the Psychological

Screening Inventory (Negy & Woods, 1993), for example, indi-

cates that the measurement of acculturation on the superficial and

intermediate levels has been useful in interpreting the results of

standardized tests. Given the complexity of the process of accul-

turation and the multiple levels on which it can be understood, the

purpose of a particular research study should inform the level of

analysis.

Although the importance of understanding the correlates and

consequences of acculturation is being increasingly recognized,

the lack of an assessment tool to measure common processes has

limited research. Because of increasing mandates from funding

agencies and professional organizations to include representative

samples of the U.S. population in research, and the recognition that

This

doc

umen

t is c

opyr

ight

ed b

y th

e A

mer

ican

Psy

chol

ogic

al A

ssoc

iatio

n or

one

of i

ts a

llied

pub

lishe

rs.

This

arti

cle

is in

tend

ed so

lely

for t

he p

erso

nal u

se o

f the

indi

vidu

al u

ser a

nd is

not

to b

e di

ssem

inat

ed b

road

ly.

MULTIGROUP ACCULTURATION SCALE 79

responsible clinical practice requires the assessment of accultura-

tion, researchers and practitioners are sorely in need of a tool to

measure the process of acculturation across ethnic groups. The

purpose of this research was the development and psychometric

evaluation of an acculturation instrument to assess some behav-

ioral and attitudinal aspects of acculturation on superficial and

intermediate levels across five ethnic groups.

Overview

Three studies were conducted to describe the development and

evaluate the psychometric properties of the Stephenson Multi-

group Acculturation Scale (SMAS). Study 1 describes the devel-

opment and refinement of the initial item pool of the SMAS using

a multiple-method process. Study 2 examined the factor structure

with exploratory factor analysis (EFA) and the internal consistency

and validity of the instrument. Study 3 examined whether the

factor structure obtained in Study 2 provided a good fit with an

independent sample using both confirmatory factor analysis (CFA)

and EFA, and it assessed convergent and discriminant validity of

the SMAS.

The construct acculturation has been plagued by conceptual

vagueness. Numerous definitions of acculturation have been pre-

sented hi the literature, most of which have been adaptations of the

most frequently used definition proposed by Redfleld, Linton, and

Herskovits (1936): "Acculturation comprehends those phenomena

which result when groups of individuals having different cultures

come into first-hand contact, with subsequent changes in the

original cultural patterns of either or both groups" (p. 149). There

are a number of discrepancies between this frequently used defi-

nition and the way in which acculturation has been actually mea-

sured. To begin with, the measurement of this construct has

focused on acculturation as an individual-level phenomenon (psy-

chological acculturation), not as a group-level or societal-level

phenomenon. Second, this definition implies the measurement of

cultural change, whereas few if any acculturation instruments

measure changes in cultural patterns. As noted earlier, accultura-

tion instruments have not tapped this process on the significant

level that involves beliefs, norms, and values that are at the heart

of culture. Although acculturation may occur at the individual,

group, and societal levels, and in the behavioral, affective, cogni-

tive, and spiritual domains, no published instrument has included

all of these levels or domains. Acculturation instruments do mea-

sure superficial-level and intermediate-level behaviors of immer-

sion or involvement in a particular society. They reflect degrees of

access of new experiences, not acquisition of the content of cul-

tural beliefs and values (Betancourt & Lopez, 1993).

In this study, acculturation is defined as degree of immersion in

dominant and elhnic societies. Degree of immersion is measured as

superficial and intermediate behaviors at the individual level in the

domains of language, interaction, food, and media. This way of

operationalizing acculturation is conducive to its application

across ethnic groups because it does not presume to measure

cultural change or the acquisition of new beliefs and values, nor to

capture the meaning of change among acculturating individuals.

As noted by Rogler (1994), acculturation, as psychologists have

been able to measure it thus far, cannot capture the meaning of

cultural change among acculturating individuals.

Study 1

The goal of Study 1 was to develop an initial item pool that was

relevant and representative of the construct acculturation. Consis-

tent with the recommendations of Haynes, Richard, and Kubany

(1995), content validation of the SMAS was a multiple-method

process in which items were derived after a review of the accul-

turation literature and previously published instruments, reviews

by expert consultants and a multiethnic research team, and field

tests with two small samples.

Method

Item development and refinement. Review of the acculturation litera-

ture and published instruments revealed common domains across instru-

ments that included language knowledge, language use, and preference;

interaction with ethnic and dominant societies; and use and preference for

foods and media (Berry & Kim, 1988; Berry & Sam, 1997; Choney,

Berryhill-Paapke, & Robbins, 1995; Cuellar et al., 1995; Landrine &

Klonoff, 1996; Marin, 1992; Marin & Gamba, 1996; Marin et al., 1987;

Mendoza, 1989; Mendoza & Martinez, 1981; (Jetting & Beauvais, 1991;

Olmedo, 1979; Padilla, 1980; Rogler et al., 1991; Suinn et al., 1987;

Szapocznik, Scopetta, Kurtines, & Aranalda, 1978; Taft, 1986; Trandis,

Kashima, Shimada, & Villareal, 1986).

After the review of the literature and previously published instruments,

an initial item pool was generated by an ethnically diverse research team

that included community professionals and consultants (N = 10). The

items were generated to reflect two independent dimensions as proposed by

Berry and Kim (1988). Scores on each of the two dimensions were

expected to measure immersion in each society. Within each dimension,

items were generated to reflect the domains of language, interaction,

media, and food. Furthermore, each domain reflected knowledge, behav-

iors, and attitudes (e.g., language knowledge, language behavior, and

language attitude). Each domain consisted of at least 30 items. The initial

pool of 195 items was then reviewed for relevance, representativeness,

specificity, accuracy, clarity, wording, and ambiguity. Item refinement

resulted in a pool of 145 items. A Likert response format was chosen with

four response options: false, partly false, partly true, and true.

Two small-scale field tests were conducted to obtain participant feed-

back and assess for completeness, clarity, suitability of format, timing,

clarity of statements, and scoring.

Participants. Participants for each phase were recruited to represent

diverse and major groups in the United States. Phase 1 included African

Americans (n = 6), participants of African descent (n — 18), Asian

Americans (n = 4), European Americans (n = 10), and Hispanic Ameri-

cans (« = 12). Phase 2 included African Americans (n = 4), participants

of African descent (n = 18), Asian Americans (n = 4), European Amer-

icans (n = 14), and Hispanic Americans (n = 10). Snowball sampling was

used because of the difficulty in identifying and recruiting some groups.

Participant's were recruited from the New York City, Boston, and Spring-

field, MA areas.

Procedure. Participants were administered the SMAS individually and

in English. After completion of the questionnaire, the participants were

debriefed and invited to comment on each item regarding relevance,

clarity, offensiveness or appearance of bias, and confusion about particular

items. In addition, they were invited to offer suggestions for possible

improvement.

On the basis of the small-scale field tests, redundant items and overall

faulty items were excluded, resulting in a pool of 115 items. These items

were then reviewed in their final form by the consultants and research

team. The final review resulted in the 95-item SMAS.

This

doc

umen

t is c

opyr

ight

ed b

y th

e A

mer

ican

Psy

chol

ogic

al A

ssoc

iatio

n or

one

of i

ts a

llied

pub

lishe

rs.

This

arti

cle

is in

tend

ed so

lely

for t

he p

erso

nal u

se o

f the

indi

vidu

al u

ser a

nd is

not

to b

e di

ssem

inat

ed b

road

ly.

80 STEPHENSON

Results

The multiple-method procedure outlined above resulted in a

95-item preliminary version of SMAS, These items were expected

to reflect two independent dimensions: dominant society immer-

sion (DSI), which consisted of 47 items and measured immersion

in dominant society, and ethnic society immersion (ESI), which

consisted of 48 items and measured immersion in one's ethnic

society. Within each dimension, items were expected to measure

the domains of language, interaction, media, and food.

Study 2

The goal of Study 2 was to examine the factor structure, internal

consistency, and construct validity of the SMAS.

Method

Participants. Participants were recruited from diverse ethnic groups,

generational statuses, ages, socioeconomic statuses (SES), and education

levels. Snowball sampling was used. The sample consisted of 436 partic-

ipants recruited from communities in the New York City area, the Boston

and Springfield areas of Massachusetts, and a large public northeastern

university. Of the sample, 62% (n = 270) were nonstudent community

participants, and 38% (n = 166) were students recruited either within these

communities or from a university setting. The age of participants ranged

from 18 to 73 years, with a mean of 29.98 years (SD = 13.3). Of the

sample, 30% (n = 132) were men and 70% women (n - 304). SES ranged

from 1 (upper middle class) to 6 (lower class), with a mean of 4.24

(SD = 1.22), as assessed by the Hollingshead (1975) index of social status.

Years of education ranged from less than seventh grade to graduate

education, with a mean of 13 years (SD = 1.5). The sample was ethnically

diverse and consisted of a number of distinct ethnic groups that were

combined to form five major groups: African Americans (n = 35, 8%);

Asian Americans (« = 33, 8%; countries of origin: Cambodia, China, Hong

Kong, India, Japan, Korea, Philippines, Thailand, Vietnam); European

Americans (n — 125,29%; countries of origin: Austria, England, Germany,

Greece, Hungary, Ireland, Italy, Poland, Portugal, Russia, Sweden); His-

panic Americans (n = 85, 19%; countries of origin: Bolivia, Brazil,

Colombia, Costa Rica, Cuba, Dominican Republic, Ecuador, Mexico, Peru,

Puerto Rico); and participants of African descent (n = 158, 36%; countries

of origin: Guyana, Haiti, Jamaica, Liberia, Trinidad). Some of the partic-

ipants who self-identified as being American of fourth generation or more

did not respond to the question pertaining to country of origin because they

did not know the country from which their ancestors came. In terms of

generational status, the sample was composed of 47.25% (n = 206) first or

immigrant generation; 19.04% (n — 83) second generation; 13.3% (n =

58) third generation; and 20.41% (n = 89) fourth generation or more. The

majority of participants identified as single (n = 271, 62%); next were

married participants (n = 108, 25%).

Measures. In addition to the SMAS, all of the participants completed

a demographic questionnaire and the Symptom Checklist-90 -Revised

(SCL-90-R; Derogatis, 1994).

The 95-item SMAS was administered to all of the participants. The items

were worded so that 47 items were designed to assess DSI and 48 were

intended to assess ESI. Items that assessed ESI used phrases such as

"native language" and "native country." Because many fourth-or-more-

generation Americans might consider English their native language or the

United States their native country, a qualifier was included at the top of the

questionnaire stating the following: "For questions that refer to native

country or country of origin, please refer to the country from which your

family originally came. For questions referring to native language, please

refer to the language spoken where your family originally came." Similar

instructions were given verbally before the questionnaire was administered

(see the Appendix).

A questionnaire administered to all participants to assess demographic

characteristics included questions about age, sex, marital status, genera-

tional status, country of origin, ethnic-racial identification, and SES (ed-

ucation and occupation).

The SCL-9Q-R (Derogatis, 1994), a frequently used standardized as-

sessment instrument, was administered to evaluate construct validity, not

as an indicator of psychopathology. Previous research studies have used

similar instruments such as the MMPI to evaluate the influence of accul-

turation on standardized test performance.

The SCL-90-R consists of ratings of 90 symptoms on a 5-point scale

indicating how frequently individuals report having experienced these

symptoms in the last week. There are nine clinical subscales and three

global scores. The most frequently used global score is the Global Severity

Index (GSI), the average score for the 90 items. Internal consistency (the

correlation of subtest scores with the total score) ranges from ,80 to .90.

Test-retest reliability coefficients range from .78 to .90. Convergent-

discriminant validity studies have indicated high correlations with inde-

pendent measures of the construct and little or no correlation with dissim-

ilar constructs (Derogatis, 1994). For the present study, the GSI score, the

average of the 90 items, was used. Cronbach's alpha for this sample

was .97.

Procedure. Two criteria were used in participant selection: over age 18

and nonrefugee status. Refugees were excluded because the traumatic

circumstances and forced nature of migration might present unique prob-

lems. Potential participants were contacted through churches, community

ethnic organization, and university ethnic organizations, and from the

general population of a large northeastern university. Participants were

contacted by telephone or in person and screened prior to instrument

administration to ensure that they met inclusion criteria. In addition,

participants were asked if their primary language was English. Those

whose primary language was not English were asked if they were fluent in

English and if they felt comfortable completing the questionnaires in

English. All of the participants included in this study serf-reported profi-

ciency in English and were administered the questionnaire in English. All

of the participants completed an informed consent form and were either

paid or given extra credit toward a university course. Questionnaires were

administered in churches, community organizations, or at the university.

Results

Exploratory factor analysis. A principal-components analysis

with varimax rotation conducted on the 95-item SMAS with the

entire sample yielded a two-factor solution. The number of factors

retained was determined by the scree test plot and eigenvalues

greater than 1.00. The two factors and their respective eigenvalues

were ESI (10.32) and DSI (5.87). Three criteria were used in the

selection of items to retain: (a) a factor loading of at least .5 on the

primary factor, (b) a difference of at least .3 between loadings on

the primary factor and loadings on the secondary factor, and (c)

correlations of less than .9 with all other items loading on the same

factor (to eliminate item redundancy). A total of 32 items were

retained that accounted for 50.6% of the scale variance; Factor 1

accounted for 27.4% and Factor 2 for 23.2%. Factor 1 included 17

items related to ESI, and Factor 2 included 15 items related to DSI.

Factor loadings and items contributing to the factors are presented

in Table 1. A large number of language items loaded heavily on

Factor 1, ESI. This is consistent with most published acculturation

scales in which items dealing with language tend to account for

much of the variance. Previous research on group-specific scales

has demonstrated a high predictive value of items dealing with

language use and preference (Marin, 1992).

This

doc

umen

t is c

opyr

ight

ed b

y th

e A

mer

ican

Psy

chol

ogic

al A

ssoc

iatio

n or

one

of i

ts a

llied

pub

lishe

rs.

This

arti

cle

is in

tend

ed so

lely

for t

he p

erso

nal u

se o

f the

indi

vidu

al u

ser a

nd is

not

to b

e di

ssem

inat

ed b

road

ly.

MULTIGROUP ACCULTURATION SCALE 81

Table 1

Items Contributing to Factors and EFA Loadings (N = 436)

Items Factor 1 Factor 2

1. I know how to speak my native language. .881 —.1162. I like to speak my native language. .834 —.129

3. I speak my native language with my friends and acquaintances from my country .827 —.134of origin.

4. I know how to read and write in my native language. .813 -.013• 5. I feel comfortable speaking my native language. .789 -.106

6. I speak my native language at home. .786 -.1977. I like to listen to music of my ethnic group. .707 —.0788. I speak my native language with my spouse or partner. .692 —.1839. When I pray, I use my native language. .678 -.225

10. I have never learned to speak the language of my native country. .677 .07211. I am informed about current affairs in my native country. .636 —.08012. I attend social functions with people from my native country. .631 —.10213. I am familiar with the history of my native country. .624 .02214. I think in my native language. .616 .06315. I stay in close contact with family members and relatives in my native country. .607 —.29916. I regularly read magazines of my ethnic group. .581 —.10417. I eat traditional foods from my native culture. .519 —.017

18. I attend social functions with (Anglo) American people. —.046 .81219. I have many (Anglo) American acquaintances. -.142 .78920. I speak English at home. -.236 .79721. I know how to prepare (Anglo) American foods. -.046 .78322. I am familiar with important people in American history. .004 .75423. I think in English. -.197 .75124. I speak English with my spouse or partner. —.230 .712

25. I feel totally comfortable with (Anglo) American people. —.085 .65926. I understand English, but I'm not fluent in English. -.180 .632

27. I am informed about current affairs in the United States. .122 .61028. I like to eat American foods. -.023 .601

' 29. I regularly read an American newspaper. .042 .59630. I feel comfortable speaking English. -.117 .59131. I feel at home in the United States. . -.085 .58632. I feel accepted by (Anglo) Americans. —.114 .554

Note. Decimals in boldface are primary factor loadings. EFA = exploratory factor analysis.

Reliability. Coefficient alphas were .86 for the entire scale observed among the first, second, and third generations did not

(Factors 1 and 2 combined) and .97 and .90 for Factors 1 and 2, hold for the fourth generation. Rather, the fourth generation means

respectively. Item total correlations ranged from .51 to .87 on were in the opposite direction than expected. The fourth generation

Factor 1 and .57 to .83 on Factor 2. scored higher on the ESI and lower on the DSI than the third

Validity. There is a strong theoretical basis for predicting a generation; however, these differences were not statistically

relationship between generational status and performance on the significant.

DSI and ESI scales (Magana et al., 1996). Previous research on Previous research findings suggested that acculturation level

individual groups suggests that, with successive generations, cer- impacts assessment results (Dana, 1993; Montgomery et a]., 1990).

tain customs of the dominant society are acquired and certain Montgomery et al. reported that when acculturation was covaried,

ethnic customs are relinquished (Cuellar et al., 1995; Keitel, Ko- significant group differences between ethnic minority and ethnic

pola, & Adamson, 1995; Marin et al., 1987; Mendoza, 1989; majority groups were reduced on the MMPI. Thus, similar patterns

Szapocznik et al., 1978). Means were computed for each item by would be expected on the SMAS, with acculturation serving as an

generation. A consistent pattern supported previous research find- explanatory mechanism in the relation between ethnic group and

ings. With each of the first three successive generations, DSI the GSI scores of the SCL-90-R. To evaluate this prediction, the

increased and ESI decreased. Item means and standard deviations current sample was divided into two groups. Group 1 consisted of

are presented in Table 2. participants who self-identified as ethnic minority group individ-

Two one-way between-groups analyses of variance indicated uals (n = 265), and Group 2 consisted of those who self-identified

that the mean differences between generations were significant on as ethnic majority group individuals (n = 125). In all, 46 cases

both the DSI, 7^(3, 432) = 73.644, p < .001, and the ESI, F(3, were deleted because of missing data, yielding a sample of 390.

432) = 31.476, p < .001. Tukey post hoc tests indicated that all Both ESI and DSI were significantly correlated with ethnic group

pairwise comparisons were significant at p < .05, except for the affiliation (r = A6,p < .001; DSI) and (r = -.39,p < .001; ESI),

third and fourth generations on both the DSI and ESI, suggesting Path analytic techniques were used to determine whether ESI and

that the greatest amount of difference occurred from the first DSI were possible mediators of the relation between ethnic group

through the third generations. The consistent pattern of means and GSI scores following the procedure presented by Baron and

This

doc

umen

t is c

opyr

ight

ed b

y th

e A

mer

ican

Psy

chol

ogic

al A

ssoc

iatio

n or

one

of i

ts a

llied

pub

lishe

rs.

This

arti

cle

is in

tend

ed so

lely

for t

he p

erso

nal u

se o

f the

indi

vidu

al u

ser a

nd is

not

to b

e di

ssem

inat

ed b

road

ly.

82 STEPHENSON

Table 2

Means and Standard Deviations by Generation

Generation

Item

no.

1234

56789

1011121314151617

Overall

1 (N

M

3.653.563.433.293.483.313.582.92

2.993.42

2.923.223.232.49

3.052.403.663.21

= 207)

SD

0.870.860.991.00

0.971.000.881.20

1.101.10

1.001.000.931.301.101.100.80

0.66

2 (N = 82)

M

Factor 1:

3.283.302.982.833.073.083.232.592.263.32

2.553.162.952.162.671.933.72

2.89

SO

Ethnic Society

1.151.071.191.211.181.121.061.36

1.281.13

1.011.000.9551.291.161.040.6130.72

3{iV =

M

Immersion

2.292.44

2.072.22

2.442.122.531.982.072.512.172.312.561.941.74

1.563.20

2.23

58)

SD

1.331.34

1.311.37

1.371.391.261.36

1.431.341.091.251.091.341.131.021.000.87

4(N =

M

2.562.752.522.58

2.672.592.832.562.572.532.112.542.662.501.672.172.93

2.39

89)

SD

1.441.42

1.461.451.441.48

1.201.471.461.441.201.341.131.481.211.371.12

0.95

Factor 2: Dominant Society Immersion

181920212223242526272829303132

Overall

3.213.032.743.163.192.952.823.062.913.153.242.953.113.013.183.05

1.081.131.181.041.011.121.240.9601.241.010.8741.111.151.030.8620.69

3.653.663.633.613.653.533.503.46

3.833.50

3.753.283.72

3.743.42

3.60

0.7770.6770.6390.7970.6520.9320.9500.804

0.5380.675

0.5110.8930.7580.6080.7720.41

3.943.983.983.963.873.983.963.87

3.983.603.873.414.00

3.893.91

3.89

0.2920.1310.1310.184

0.4890.131

0.5670.3780.1310.5280.4980.9920.000

1.0300.8620.12

3.973.953.943.91

3.843.893.893.783.883.593.913.613.98

3.933.713.87

0.1500.2570.3470.3880.4220.4530.1840.596

0.5130.5580.3880.8330.106

0.3630.624

1.80

Kenny (1986). DSI scores were regressed on ethnic group, F(l,

388) = 118.68, p < .001; GSI scores were regressed on ethnic

group, F(l, 388) = 8.637, p = .003; and GSI scores were re-

gressed on both ethnic group and DSI. Results indicated that DSI

mediated the effects of ethnic group on GSI scores when ethnic

group and DSI were controlled. The previously significant relation

between ethnic group and GSI scores was no longer significant.

The resulting weights presented in Figure 1 provide estimates of

the effects of the variables. The same procedure was followed with

the ESI scores. ESI scores were regressed on ethnic group, F(l,

388) = 78.66, p s .001. Regression of GSI scores on both ethnic

group and ESI indicated that ESI did not mediate the effects of

ethnic group on GSI scores. Consistent with previous research,

DSI mediated the effects of ethnic group affiliation and assessment

results.

Study 3

The goal of Study 3 was to evaluate whether the factor structure

obtained in the Study 2 EFA was robust across samples and

provided a good fit with a new sample and to evaluate the con-

vergent and discriminant validity of the SMAS in relation to two

other acculturation instruments.

Method

Participants. The sample for the second study consisted of 208 under-

graduate students enrolled at a large northeastern university. The age of

participants ranged from 18 to 60 years, with a mean of 22.8 years

(SD = 7.32). Of the participants, 21% (n = 43) were men and 79% (n =

165) women. In terms of generational status, 14% (n = 29) were first

generation, 17% (n — 36) were second generation, 18% (n = 38) were

third generation, and 51% (n = 105) were fourth generation or more. SES

This

doc

umen

t is c

opyr

ight

ed b

y th

e A

mer

ican

Psy

chol

ogic

al A

ssoc

iatio

n or

one

of i

ts a

llied

pub

lishe

rs.

This

arti

cle

is in

tend

ed so

lely

for t

he p

erso

nal u

se o

f the

indi

vidu

al u

ser a

nd is

not

to b

e di

ssem

inat

ed b

road

ly.

MULTIGROUP ACCULTURATION SCALE 83

DSI

0.66'

Ethnic Group1

5.41*

GSI

0.58 (ns)

Figure 1. Regression coefficients estimating the mediating effect of

dominant society immersion (DSI) on the relation between ethnic group

affiliation and Global Severity Index (GSI) scores. *p < .001.

ranged from 1 (upper middle class) to 6 (lower class), witii a mean of 5

(SD = 1.41). In terms of ethnic background, the sample included 3% (n =

7) participants of African descent, 15% (n = 31) Hispanic Americans, 73%

(n = 151) European American, 7% (n = 15) Asian Americans, and 2%

(n = 4) African American. Years of education ranged from high school

graduate to graduate studies, with a mean education of 14 years (SD = .85).

Measures. All of the participants completed the SMAS, the demo-

graphic questionnaire, and two bidimensiona! acculturational scales. The

Acculturation Rating Scale for Mexican Americans-II (ARSMA-II; Cuellar

et al., 1995) and the Bidimensional Acculturation Scale for Hispanics

(BAS; Marin & Gamba, 1996) were used in this study for several reasons.

To begin with, there are no published instruments for any of the ethnic

groups included in this study, except for Hispanic Americans, that are

consistent with the most recent conceptualizations of the acculturation

process. Both the ARSMA-II and the BAS are widely used and reliable

instruments that measure acculturation through an orthogonal, bidimen-

sional approach. Second, both scales reflected the operationalization of

acculturation as degree of immersion, measured as superficial and inter-

mediate behaviors at the individual level. Both scales were modified for

use with diverse ethnic groups as described below.

The ARSMA-II is a 50-item multidimensional acculturation scale de-

signed to measure language use and preference, ethnic identity and clas-

sification, cultural heritage and cultural behaviors, and ethnic interaction in

Mexican Americans. The development sample included 379 university

students representing five generations. The first generation represented

individuals born in Mexico, and subsequent generations were assessed on

the basis of parent's and grandparent's place of birth. The sample included

groups from varying SES statuses as assessed with the Duncan Socioeco-

nomic Index. The ARSMA-II consists of two subscales. Scale 1 measures

Mexican orientation (MOS) and Anglo orientation (AOS), and Scale 2 (the

Marginality scale) measures separation and marginalization acculturation

modes. Scale 2, the Marginality scale, is considered an experimental scale,

inasmuch as adequate validity has not yet been established. Only Scale 1

was used for this study. Cuellar et al. (1995) recommended the use of

Scale 1 independent of Scale 2. They consider Scale 2 independent of the

ARSMA-II instrument and suggest that either Scale 1 or Scale 2 can be

administered, scored, and interpreted without the other. Scale 1 consists

of 30 items, 17 MOS items and 13 AOS items. Sample MOS and AOS

items are "I associate with Mexicans and or Mexican Americans" and "My

friends are now of Anglo origin." The scale possesses good internal and

test-retest reliability, with coefficients alpha of .83 and .88 for the AOS

and MOS, respectively. Test-retest correlations over a 1-week interval

were .94 and .96 for the AOS and MOS, respectively (Cuellar et al., 1995).

The BAS is a 24-item scale developed for use with Mexican Americans

and Central Americans. The development sample included 254 adults,

most of whom were first-generation participants born outside the United

States (80%), with a mean of 10.4 years of formal education. The BAS

measures language-related aspects of acculturation in both ethnic (His-

panic) and dominant (non-Hispanic) societies. The instrument has high

internal consistency, with an alpha of .94 for the Non-Hispanic domain and

.87 for the Hispanic domain (Marin & Gamba, 1996). Sample Hispanic

domain and Non-Hispanic domain BAS items are "How often do you

speak Spanish?' and "How often do you speak English?"

Both the ARSMA-II and the BAS were modified for use with diverse

ethnic groups. For example, Item 1 on the ARSMA-II, "I speak Spanish,"

was modified to read "I speak my native language." Item 6 on the BAS,

"How often do you speak Spanish?" was modified to read "How often do

you speak your native language?" All Hispanic domain items on the BAS

and all MOS items on the ARSMA-H were modified for use with diverse

ethnic groups. Items assessing immersion in dominant society, the AOS on

the ARSMA-n and the Non-Hispanic domain on the BAS, were not

modified Cronbach's alpha for this sample with the modified ARSMA-n

was .86, and .92 with the modified BAS.

Procedure. The same criteria were used for participant selection as in

Study 2. Potential participants were recruited from a general university

population. Participants were screened prior to instrument administration

and completed informed consent forms. AH of the participants were given

extra credit toward a university course. Questionnaires were administered

at the university.

Confirmatory factor analysis. On the basis of the factor structure

obtained in Study 2 with EFA, a model was generated and tested with CFA

to evaluate its validity given the Study 3 data. The CFA model was

estimated with the personal computer version of LISREL 8 (JSreskog &

Sdrbom, 1993). An independence model hypothesized a priori that (a)

responses on the SMAS could be explained by two factors, ESI and DSI,

(b) variables would have a nonzero loading on the factor it was designed

to measure and a zero loading on the other factor, (c) the two /actors would

be uncorrelated, and (d) measurement error terms would be uncorrelated.

The recommended indexes and criteria used to determine the fit of the

CFA model were based on the recommendations of Hoyle and Panter

(1995) and included two absolute indexes (chi-square and the goodness of

fit index [GFI]), one Type-2 index, (the incremental fit index [IFI]), and

one Type-3 index (the comparative fit index [CFI]). Criteria included a

chi-square to degree of freedom ratio of about 2:1; CFI (Bentler, 1990),

GFI (Joreskog & Sorbom, 1989), and IFI (Bollen, 1989) of .90 or greater.

The .90 cutoff stands as the agreed upon cutoff for overall fit indexes

(Hoyle, 1995; Hoyle & Panter, 1995).

Because of the categorical nature of the data, the large number of

individual items, and the general nonnormality of the data, as indicated by

skewness and kurtosis statistics, several steps were taken before (he data

was submitted for analysis. A method of reexpression was used to produce

item parcels with fewer parameters that would eliminate redundancies and

produce a distribution that more closely approximated normality (Floyd &

Widamam, 1995; West, Finch, & Curran, 1995). The 32 individual items

of the SMAS were combined to create parcels that reflected specific

theory-driven domains. Thus, 10 parcels were created and 5 parcels spec-

ified to load on each of the two factors. Language 1 = 10 items, Interaction

1 = 2 items, Knowledge 1 = 2 items, Media 1 = 2 items, and Food 1 = 1

item were specified to load on Factor 1. Language 2 = 5 items, Interaction

2 - 5 items, Knowledge 2 = 2 items, Media 2 = 1 item, and Food 2 = 2

items were specified to load on Factor 2. In addition, Prelis 2 (Joreskog &

Sorbom, 1993) was used to structure a polychoric correlation matrix

recommended for ordered categorical variables (Joreskog & Sorbom, 1993;

Wothke, 1995) from the raw1 data and listwise deletion of missing data

(Wothke, 1995; N = 187). The entire correlation matrix for the SMAS

variables is available from the author upon request.

Results of the confirmatory factor analysis. Results of the CFA suggest

inconsistent findings. In terms of the chi-square GFI test, the model did not

prove consistent with the observed data; ̂ (34, N - 187) = 89.89, p = <

.001. In terms of adjunct fit statistics, the model provided a close approx-

imation to the observed data; GFI = .91, IFI = .90, and CFI = .90.

This

doc

umen

t is c

opyr

ight

ed b

y th

e A

mer

ican

Psy

chol

ogic

al A

ssoc

iatio

n or

one

of i

ts a

llied

pub

lishe

rs.

This

arti

cle

is in

tend

ed so

lely

for t

he p

erso

nal u

se o

f the

indi

vidu

al u

ser a

nd is

not

to b

e di

ssem

inat

ed b

road

ly.

84 STEPHENSON

Examination of the standardized residuals indicated a poor fit in the

prespecified uncorrelated error terms between the following parcels: Lan-

guage 1 and Media 1; 4.29, Media 2 and Knowledge 1: 4.20, and Knowl-

edge 1 and Knowledge 2: 3.22. The large positive standardized residuals

indicated that the model underestimated the covariance between these sets

of variables and that two of the correlated error terms occurred across two

different factors. West et al. (1995) reported that results of simulation

studies indicated that as categorized variables become increasingly skewed,

the chi-square values become inflated, and correlations may be spuriously

obtained between error variances associated with items having similar

degrees of skewness. An item with similar agreement rates, for example,

can give rise to spurious correlations reflecting only the common degree of

skewness among items. The measures taken before data submission may

not have been sufficient to remediate the degree of nonnormality in the data

and meet the stringent criteria for obtaining a reasonable fit based on

chi-square statistics. However, inspection of the CFI and the IFI recom-

mended for smaller sample sizes (<200) and that produce only a small

downward bias even under severely nonnormal condition (West et al.,

1995) indicated that both CFI and IFI reached the agreed upon cutoff of .90

or greater.

The parcels loaded positively on their corresponding prespecified fac-

tors, indicating that the two-factor structure does explain the observed

interrelations among the SMAS items. The findings also suggested that the

two factors are correlated, with an interfactor correlation of —.32. As

reported by West et al. (1995), the underestimated standard errors, as

indicated by the large positive residuals, may have produced correlations

between factors even if they do not exist in the population.

Exploratory factor analysis with Study Sample 3, Because it was

difficult to determine whether the CFA model provided a plausible repre-

sentation of the structure of the observed data, EFA was conducted with the

Study 3 sample and the EFA solutions from Studies 2 and 3 were compared

to assess for similarity of factor loadings across samples.

A principal-components analysis using a varimax rotation was per-

formed to examine the factor structure of the items with the second sample.

Analysis yielded a two-factor solution with eigenvalues of 10.05 for the

ESI and 4.34 for the DSL Every item loaded on the same factor as in

Study 2. However, the criterion of loadings of at least .5, was not met on

eight items. Three items on the ESI did not meet this retention criterion;

their loadings ranged from .34 to .49. Similarly, five items failed to meet

this same criterion on the DSI, and ranged from .34 to .49. Also, three items

failed to meet the criterion of a difference of at least .3 between loadings

on the primary factor and loadings on the secondary factor. Factor 1 (ESI)

accounted for 28% of the variance, and Factor 2 (DSI) accounted for 17%

(total 45%). Factor loadings with the Study 3 sample were not as strong as

those found in Study 2, and the variance accounted for by the factors was

comparable for Factor 1 (ESI; 27% and 28%) and somewhat less for

Factor 2 (DSI; 23% and 17%). The difference in the variance accounted for

by Factor 1 (DSI) may reflect the restricted range on the DSI scores of the

sample in Study 3. Of the Study 3 sample, 51% was composed of partic-

ipants from the fourth generation, compared with 20% in the Study 2

sample, indicating either that the DSI items were functioning as intended

or that the items were not sensitive enough to discriminate differences in

latter generations, or both. Table 3 presents the results of the factor

analysis.

Assessment of similarity of factor loadings across samples. The salient

variable similarity (s) index (Cattell, Balcar, Horn, & Nesselroade, 1969)

was used to determine whether factor structure was similar across the two

samples. This involved generating matrices of interfactorial similarity,

placing these comparisons into a 3 X 3 contingency table depending on

positive salience, hyperplane, or negative salience, and calculating the

index using the formula presented by Cattell et al. In previous research,

hyperplane cutoff levels have been set from . 10 to .40 (Cattell et al., 1969;

McCormick, Green, & Walkey, 1987). Hyperplane cutoff level was spec-

ified at a conservative level of .10. Results indicated a similarity index ,82

for the ESI and .78 for the DSL Significance of the s values was calculated

using probability tables generated by Cattell et al. Results indicated that the

factor structure was highly similar across samples.

Reliability. Coefficient alphas were .94 and .75 for Factors 1 and 2,

respectively.

Validational analysis. Correlational studies were conducted to evalu-

ate the validity of the SMAS by examining the relation between the two

SMAS subscales and those of two other acculturation instruments; the

ARSMA-n and the BAS. Results suggest that the ESI subscale of the

SMAS was strongly correlated with the MOS of the ARSMA-H (r = .87,

p = < .01) and negatively correlated with the AOS (r = -.28, p < .01).

The ESI was positively correlated with the Hispanic Domain scale of the

BAS (r = .83, p < .01) and negatively correlated with the Non-Hispanic

Domain scale (r = ~.25,p < .01).

The DSI subscale of the SMAS was positively correlated with the AOS

(r = .49, p < .01) and negatively correlated, although not significantly,

with the MOS (r = -.I5,p = nj). With the BAS, the DSI was positively

correlated with the Non-Hispanic scale (r = .48, p < .01) and negatively

correlated, although not significantly, with the Hispanic scale (r = —.17,

p — ns). The ESI was significantly correlated in the expected direction with

both the ARSMA-II and the BAS. Similarly, the DSI was correlated in the

expected direction with both instruments.

General Discussion

The SMAS is the first acculturation scale developed for use

across ethnic groups. The catalyst for the development of this

Table 3

EFA Loadings for Study 3 Sample (N = 205;

Item

6

21

3145498

10167

1211

151317202318212530

193231282427222926

Factor 1

.887

.872

.870

.849

.848

.828

.827

.809

.773

.747

.636

.598

.541

.502

.486

.435

.335-.165-.256-.116-.123-.105-.014-.064-.208-.127

.131-.100

.001-.076

.046-.210

Factor 2

-.105-.062-.134-.098

.017-.087-.138

.047

-.027-.121-.095-.148-.134-.115-.345-.135-.129

.772

.717

.705

.650

.603

.594

.587

.584

.569

.537

.494

.480

.467

.449

.335

Note. Items are numbered to reflect item numbers of original factorloadings with the entire sample. Decimals in boldface are primary factorloadings. EFA ~ exploratory factor analysis.

This

doc

umen

t is c

opyr

ight

ed b

y th

e A

mer

ican

Psy

chol

ogic

al A

ssoc

iatio

n or

one

of i

ts a

llied

pub

lishe

rs.

This

arti

cle

is in

tend

ed so

lely

for t

he p

erso

nal u

se o

f the

indi

vidu

al u

ser a

nd is

not

to b

e di

ssem

inat

ed b

road

ly.

MULT1GROUP ACCULTURATION SCALE 85

instrument was the need for a reliable and valid instrument that can

be used across groups to facilitate widespread assessment of this

important construct. These preliminary data indicate that the

SMAS holds promise as a valuable tool. The scale does not

measure all possible areas related to acculturation, nor does it

measure beliefs, norms, and values. However, this instrument does

provide an index of degree of immersion in both dominant and

ethnic societies that can facilitate interpretation of research, as-

sessment data, and clinical presentation.

EFA findings in Study 2 suggested a two-factor solution of the

SMAS consistent with the theoretical conceptualization of the

acculturation process. CFA conducted in Study 3 did not show the

model to be consistent with the observed data based on chi-square

statistics, although adjunct GFIs did reach adequate levels to

indicate acceptable fit. The factor structure that resulted from the

EFA in Study 3 was consistent with that found in Study 2 and

proved robust across samples, despite the differences between the

two samples. The reliability and validity studies indicated nigh

reliability and validity indexes comparable with other published

instruments. Strong validity was demonstrated in the instrument's

ability to identify relationships between generational status and

performance on the subscales. Additional support for the validity

of the SMAS and the importance of assessing acculturation was

the finding that the DSI subscale served as a mediator that cor-

rected for differences in the GSI of the SCL-90-R between ethnic

majority and ethnic minority groups. These findings strongly sup-

port the argument that assessing acculturation is a valuable tool in

making sense of between-group differences without resorting to

deficit explanations. These findings underline the argument that

individuals who are less immersed in dominant society may re-

quire alternative culturally sensitive assessment before test results

or clinical presentation can be understood or interpreted.

These results are tentative, and replication with larger represen-

tative samples are needed before definitive conclusions are drawn.

One limitation of these studies is that the results are based on data

from nonrandom samples and cannot be generalized to other

groups. Also, because this instrument has not yet been adapted into

other languages, all of the participants were administered the

questionnaires in English. This limitation may contribute to some

bias, particularly because of the large numbers of strongly loaded

language items. Despite the fact that all of the participants com-

pleted the questionnaire in English, the instrument was able to

detect separated and marginalized individuals. The data suggest

that 16% of participants were either separated or marginal.

Although it was expected that participants of the third and fourth

generations would be more immersed in dominant society than

participants from the first and second generations, it is difficult to

determine whether the restricted range of scores observed on the

DSI with latter generations suggests that the SMAS items are

functioning as intended or suggests potential limitations in assess-

ing latter generations. Also, the large numbers of language items

and references to country of origin, coupled with the fact that many

fourth-generation participants did not know the countries from

which their families originally came, suggest the need to interpret

the findings of the ESI subscale with caution with fourth-

generation participants. As such, acculturation as measured in this

study may be more useful for newer immigrants of the first and

second generations.

The fact that previous research has found that fourth-generation

minority individuals (e.g., African Americans) perform differently

than their nonminority (e.g., European Americans) counterparts on

some standardized instruments indicates that attention to accultur-

ation experiences beyond the third generation should be an impor-

tant area of future research. Future research might assess accul-

turation in subsequent generations on the more significant level

involving beliefs, norms, and values rather than superficial and

intermediate involvement and acquisition of new experiences in a

given society. Superficial and intermediate levels of involvement

and acquisition of experiences hi dominant society would have

already occurred in the fourth generation. Research might further

investigate language and country of origin items that tend to

account for much of the variance in acculturation instruments. For

example, in some cases, English is considered the language of

origin because latter generation immigrants do not know the coun-

try from which their family came. Language items could include

participants' knowledge and use of other forms of English, such as

nonstandard American English and dialects for European Ameri-

cans, dialects and Creoles for individuals whose ancestors immi-

grated from former English colonies, or ebonies for African Amer-

icans. Interaction items might include contact and interactions with

their respective ethnic groups in the United States rather than in the

country of origin.

Many conceptual and methodological issues remain in the mea-

surement of this important and complex construct that merit re-

search attention. The findings of these present studies have under-

lined the importance of assessing acculturation and its usefulness

as a tool for understanding between-group differences. The overall

robust outcome of these studies indicated that the SMAS does

provide an initial index of degree of immersion hi both dominant

and ethnic societies that can serve as a valuable tool in facilitating

interpretation of research, assessment data, and clinical presenta-

tion, particularly with more recent immigrants to the United States.

References

Arnold, B. R., Montgomery, G. T., Castaneda, I., & Longoria, R. (1994).

Acculturation and performance of Hispanics on selected Halstead-

Reitan neurological tests. Assessment, 1, 239-248.

Atkinson, D. R., Casas, A., & Abreu, J. (1992). Mexican-American accul-

turation, counselor ethnicity and cultural sensitivity, and perceived coun-

selor preference. Journal of Counseling Psychology, 39, 515-520.

Atkinson, D. R., Thompson, C. E., & Grant, S. K. (1993). A three-

dimensional model for counseling racial-ethnic minorities. The Coun-

seling Psychologist, 21, 257-277.

Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable

distinction in social psychological research: Conceptual, strategic, and

statistical consideration. Journal of Personality and Social Psychol-

ogy, 51, 1173-1182.

Bentler, P. M. (1990). Comparative fit indexes in structural models. Psy-

chological Bulletin, 107, 238-246.

Berry, J. W. (1980). Acculturation as varieties of adaptation. In A. Padilla

(Ed.), Acculturation: Theory, models and new findings (pp. 9-25).

Boulder, CO: Westview Press.

Berry, J. W. (1992). Acculturation and adaptation in a new society.

International Migration, 30, 69-85.

Berry, J. W. (1996). Immigration, acculturation and adaptation. Applied

Psychology: An International Review, 46, 5-34.

Berry. J. W., & Kim, U. (1988). Acculturation and mental health. In P. R.

Dansen, J. W. Berry, & N. Sartorius (Eds.), Health and cross-cultural

psychology (pp. 207-236). Newberry Park, CA: Sage.

This

doc

umen

t is c

opyr

ight

ed b

y th

e A

mer

ican

Psy

chol

ogic

al A

ssoc

iatio

n or

one

of i

ts a

llied

pub

lishe

rs.

This

arti

cle

is in

tend

ed so

lely

for t

he p

erso

nal u

se o

f the

indi

vidu

al u

ser a

nd is

not

to b

e di

ssem

inat

ed b

road

ly.

STEPHENSON

Berry, J. W., & Sam, D. (1997). Acculturation and adaptation. In J. W.

Berry, M. H. Segall, & C. Kagitcibasi (Eds.), Handbook of cross-

cultural psychology. Social behavior and applications (pp. 291-326).

Boston: Allyn & Bacon.

Betancourt, H., & Lopez, S. R. (1993). The study of culture, ethnicity, and

race in American psychology. American Psychologist, 48, 629-637.

Bollen, K. A. (1989). A new incremental fit index for general structural

equation models. Sociological Methods and Research, 17, 303—316.

Cattell, R. B., Balcar, K. R., Horn, J. L., & Nesselroade, J. R. (1969).

Factor matching: An improvement of the S index; with tables. Educa-

tional and Psychological Measurement, 29, 781-792.

Choney, S. K., Berryhill-Paapke, E., & Robbins, R. R. (1995). The accul-

turation of American Indians: Developing frameworks for research and

practice. In J. G. Ponterotto, J. M. Casas, L. A. Suzuki, & C. M.

Alexander (Eds,), Handbook of multicultural counseling {pp. 73-92).

Thousand Oaks, CA: Sage.

Cuellar, L, Arnold, B., & Maldonado, R. (1995). Acculturation Rating

Scale for Mexican Americans: II. A revision of the original ARSMA

scale. Hispanic Journal of Behavioral Sciences, 17, 275-304.

Dana, R. H. (1993). Multicultural assessment perspectives for professional

psychology. Boston: Allyn & Bacon.

Dana, R. H. (1995). Impact of the use of standard psychological assessment

on the diagnosis and treatment of ethnic minorities. In J. F. Aponte,

R. Y. Rivers, & J. Wohl (Eds.), Psychological interventions and cultural

diversity (pp. 57-73). Needham Heights, MA: Allyn & Bacon.

Derogatis, L. R. (1994). Symptom Checklist-90-R: Administration, scor-

ing, and procedure manual (3rd ed.). Minneapolis, MN: National Com-

puter Systems.

Dinges, N. G., & Cherry, D. (1995). Symptom expression and the use of

mental health services among American ethnic minorities. In J. F.

Aponte, R. Y. Rivers, & J. Wohl (Eds.), Psychological interventions and

cultural diversity (pp. 40-56). Needham Heights, MA: Allyn & Bacon.

Floyd, F. J., & Widaman, K. F. (1995). Factor analysis in the development

and refinement of clinical assessment instruments. Psychological As-

sessment, 7, 286-299.

Greene, R. L. (1987). Ethnicity and MMPI performance: A review. Journal

of Consulting and Clinical Psychology, 55, 497-512.

Grieger, I., & Ponterotto, J. G. (1995). A framework for assessment in

multicultural counseling. In J. G. Ponterotto, J. M. Casas, L. A. Suzuki,

& C. M. Alexander (Eds.), Handbook of multicultural counseling (pp.

357-374). Thousand Oaks, CA: Sage.

Hathaway, S. R., & McKinley, J. C. (1943). The Minnesota Multiphasic

Personality Schedule. Minneapolis: University of Minnesota Press.

Haynes, S. N., Richard, D. C., & Kubany, E. S. (1995). Content validity in

psychological assessment: A functional approach to concepts and meth-

ods. Psychological Assessment, 7, 238-247.

Hollingshead, A. B. (1975). Four factor index of social status. New Haven,

CT: Yale University.

Hoyle, R. H. (1995). The structural equation modeling approach. Basic

concepts and fundamental issues. In R. H. Hoyle (Ed.), Structural

equation modeling: Concepts, issues and applications (pp. 1—15). Thou-

sand Oaks, CA: Sage.

Hoyle, R. H., & Panter, A. T. (1995). Writing about structural equation

models. In R, H. Hoyle (Ed.), Structural equation modeling: Concepts,

issues and applications (pp. 158-176). Thousand Oaks, CA: Sage.

Joreskog, K. G., & Sorbom, D. (1989). USREL 7: A guide to the program

and applications. Chicago: SPSS.

Joreskog, K. G., & Sorbom, D. (1993). LISREL 8: Structural equation

modeling with the Simplis command language. Chicago: Scientific Soft-

ware.

Keitel, M. A., Kopola, M., & Adamson, W. S. (1995). Ethical issues in

multicultural assessment. In L. A. Suzuki, P. J. Meller, & J. G. Pon-

terotto (Eds.), Handbook of multicultural assessment: Clinical, psycho-

logical and educational applications (pp. 29-48). San Francisco:

Jossey-Bass.

Landrine, H., & Klonoff, E. A. (1996). African American acculturation:

Deconstructing race and reviving culture. Thousand Oaks, CA: Sage.

Magana, J, R., de la Rocha, O-, Amsel, J., Magana, H. A., Fernandez, M. I.,

& Rulnick, S. (1996). Revisiting the dimensions of acculturation: Cul-

tural theory and psychometric practice. Hispanic Journal of Behavioral

Sciences, 18, 444-468.

Marin, G. (1992). Issues in the measurement of acculturation among

Hispanics. In K. F. Geisinger (Ed.), Psychological testing ofHispanics

(pp. 235-251). Washington, DC: American Psychological Association.

Marin, G., & Gamba, R. J. (1996). A new measurement of acculturation for

Hispanics: The Bidimensional Acculturation Scale for Hispanics (BAS).

Hispanic Journal of Behavioral Sciences, 18, 297-316.

Marin, G., Gamba, R. J., & Marin, B. V. (1992). Extreme response style

and acquiescence among Hispanics: The role of acculturation and edu-

cation. Journal of Cross-Cultural Psychology, 23, 498-509.

Marin, G., Sabogal, F., VanOss Marin, B., Otero-Sabogal, R., &. Perez-

Stable, E. (1987). Development of a short acculturation for Hispanics.

Hispanic Journal of Behavioral Sciences, 9, 183-205.

Matsumoto, G. M., Meredith, G. M., & Masuda, M. (1970). Ethnic iden-

tification: Honolulu and Seattle Japanese-Americans. Journal of Cross-

Cultural Psychology, 1, 63-76.

McCormick, I. A., Green, D. E., & Walkey, F. H. (1987). A multiple

replication of the three factor solution of the Eysenck Personality In-

ventory identified using a two-step confirmation and two factor replica-

tion procedure, FACTOREP, Journal of Personality and Individual

Differences, 8, 285-287.

Mendoza, R. H. (1989). An empirical scale to measure type and degree of

acculturation in Mexican-American adolescents and adults. Journal of

Cross-Cultural Psychology, 20, 373-385.

Mendoza, R. H., & Martinez, J. L. (1981). The measurement of accultur-

ation. In A. Baron (Ed.), Explorations in Chicano psychology (pp.

71-82). New York: Praeger.

Montgomery, G. T. (1992). Acculturation, stressors, and somatization

patterns among students from extreme south Texas. Hispanic Journal of

Behavioral Sciences, 14, 434-454.

Montgomery, G. T., Arnold, B., & Orozco, S. (1990). MMPI supplemental

scale performance of Mexican Americans and level of acculturation.

Journal of Personality Assessment, 54, 328-342.

Montgomery, G. T., & Orozco, S. (1985). Mexican Americans' perfor-

mance on the MMPI as a function of level of acculturation. Journal of

Clinical Psychology, 41, 203-212.

Moreland, K. L. (1996). Persistent issues in multicultural assessment of

social and emotional functioning. In L. A. Suzuki, P. J. Meller, & J, G.

Ponterotto (Eds.), Handbook of multicultural assessment: Clinical, psy-

chological and educational applications (pp. 51-76). San Francisco:

Jossey-Bass.

Negy, E. R., & Woods, D. J. (1993). Mexican- and Anglo-American

differences on the Psychological Screening Inventory. Journal of Per-

sonality Assessment, 60, 543-553.

Getting, E. R., & Beauvais, F. (1990). Orthogonal cultural identification

theory: The cultural identification of minority adolescents. International

Journal of Addictions, 25, 655-685.

Olmedo, E. L. (1979). Acculturation: A psychometric perspective. Amer-

ican Psychologist, 34, 1061-1070.

Olmedo, E. L., Martinez, J. L., & Martinez, S. R. (1979). Measure of

acculturation for Chicano adolescents. Psychological Reports. 42, 159-

170.

Padilla, A. M. (1980). Acculturation: Theory, models and some new

findings. Boulder, CO: Westview Press.

Redfield, R., Union, R., & Herskovits, M. J. (1936). Memorandum for the

study of acculturation. American Anthropologist, 38, 149-152.

This

doc

umen

t is c

opyr

ight

ed b

y th

e A

mer

ican

Psy

chol

ogic

al A

ssoc

iatio

n or

one

of i

ts a

llied

pub

lishe

rs.

This

arti

cle

is in

tend

ed so

lely

for t

he p

erso

nal u

se o

f the

indi

vidu

al u

ser a

nd is

not

to b

e di

ssem

inat

ed b

road

ly.

MULTIGROUP ACCULTURATION SCALE 87

Rogler, L. H. (1994). International migration: A framework for directing

research. American Psychologist, 49, 701-708.

Rogler, L. H., Cortes, D. E., & Malgady, R. G. (1991). Acculturation and

mental health among Hispanics. American Psychologist, 585-597.

Sodowsky, G. R., & Carey, J. C. (1988). Relationships between

acculturation-related demographics and cultural attitudes of an Asian-

Indian immigrant group. Journal of Multicultural Counseling and De-

velopment, 16, 117-136.

Suinn, R. M, Rickard-Figueroa, K., Lew, S., & Vigil, P. (1987). The

Suinn-Lew Asian Self-Identification Acculturation Scale: An initial re-

port. Educational and Psychological Measurement, 47, 401—407.

Suzuki, L. A., Vraniak, D. A., & Kugler, J. F. (1996). Intellectual assess-

ment across cultures. In J. G. Ponterotto, J. M. Casas, L. A. Suzuki, &

C. M. Alexander (Eds.), Handbook of multicultural counseling (pp.

141-177). Thousand Oaks, CA: Sage.

Szapocznik, J., & Kurtines, W. (1993). Family psychology and cultural

diversity: Opportunities for theory, research and applications. American

Psychologist, 48, 400-407.

Szapocznik, J., Scopetta, M. A., Kurtines, W., & Aranalda, M. (1978).

Theory and measurement of acculturation. Interamerican Journal of

Psychology, 12, 113-130.

Taft, R. (1986). The psychological study of the adjustment and adaptation

of immigrants in Australia. Australian Journal of Psychology, 38, 364-

386.

Trandis, H. C., Kashima, Y., Shimada, E., & Villareal, M. (1986). Accul-

turation indices as a means of confirming cultural differences. Interna-

tional Journal of Psychology, 21, 43-70.

Trickett, E. J., Watts, R. J., & Birman, D. (1994). Toward an overarching

framework for diversity. In E. I. Trickett, R. J. Watts, & D. Birman

(Eds.), Human diversity: Perspectives on people in context (pp. 7-26).

San Francisco: Jossey-Bass.

Watts, R. J. (1994). Paradigms of diversity. In E. J. Trickett, R. J. Watts,

& D. Birman (Eds.), Human diversity: Perspectives on people in context

(pp. 49-80). San Francisco: Jossey-Bass.

West, S. G., Finch, J. F., & Curran, P. J. (1995). Structural equation models

with non normal variables: Problems and remedies. In R. H. Hoyle (Ed.),

Structural equation modeling: Concepts, issues and applications (pp.

56-75). Thousand Oaks, CA: Sage.

Wothke, W. (1995). Nonpositive definite matrices in structural modeling.

In R. H. Hoyle (Ed.), Structural equation modeling: Concepts, issues

and applications (pp. 256-293). Thousand Oaks, CA: Sage.

Yao, E. L. (1979). The assimilation of contemporary Chinese immigrants.

Journal of Psychology, 101, 107-113.

Zalewski, C., & Greene, R. L. (1996). Multicultural usage of the MMPI-2.

In J. G. Ponterotto, I. M. Casas, L. A. Suzuki, & C. M. Alexander (Eds.),

Handbook of multicultural counseling (pp. 77-114). Thousand Oaks,

CA: Sage.

(Appendix follows)

This

doc

umen

t is c

opyr

ight

ed b

y th

e A

mer

ican

Psy

chol

ogic

al A

ssoc

iatio

n or

one

of i

ts a

llied

pub

lishe

rs.

This

arti

cle

is in

tend

ed so

lely

for t

he p

erso

nal u

se o

f the

indi

vidu

al u

ser a

nd is

not

to b

e di

ssem

inat

ed b

road

ly.

STEPHENSON

Appendix

Stephenson Multigroup Acculturation Scale (SMAS)

Below are a number of statements that evaluate changes that occur when people interact with others of different cultures

or ethnic groups. For questions that refer to "COUNTRY OF ORIGIN" or "NATIVE COUNTRY," please refer to the

country from which your family originally came. For questions referring to "NATIVE LANGUAGE," please refer to the

language spoken where your family originally came.

Circle the answer that best matches your response to each statement

False Partly false Partly true True

1. I understand English, but I'm not fluent in English.

2. I am informed about current affairs in the United States.

3. I speak my native language with my friends and acquaintances from my country of origin.

4. I have never learned to speak the language of my native country.

5. I feel totally comfortable with (Anglo) American people.

6. I eat traditional foods from my native culture.

7. I have many (Anglo) American acquaintances.

8. I feel comfortable speaking my native language.

9. 1 am informed about current affairs in my native country.

10. I know how to read and write in my native language.

11. I feel at home in the United States.

12. I attend social functions with people from my native country.

13. I feel accepted by (Anglo) Americans.

14. I speak my native language at home.

15. I regularly read magazines of my ethnic group.

16. I know how to speak my native language.

17. I know how to prepare (Anglo) American foods.

18. I am familiar with the history of my native country.

19. I regularly read an American newspaper.

20. I like to listen to music of my ethnic group.

21. I like to speak my native language.

22. I feel comfortable speaking English.

23. I speak English at home.

24. I speak my native language with my spouse or partner.

25. When I pray, I use my native language.

26. I attend social functions with (Anglo) American people.

27. I think in my native language.

28. I stay in close contact with family members and relatives in my native country.

29. I am familiar with important people in American history.

30. I think in English.

31. I speak English with my spouse or partner.

32. I like to eat American foods.

Copyright (1998) by Margaret Stephenson. This instrument may be reproduced with permission from Margaret Stephenson.

Received March 30, 1999

Revision received November 5, 1999

Accepted November 11, 1999

This

doc

umen

t is c

opyr

ight

ed b

y th

e A

mer

ican

Psy

chol

ogic

al A

ssoc

iatio

n or

one

of i

ts a

llied

pub

lishe

rs.

This

arti

cle

is in

tend

ed so

lely

for t

he p

erso

nal u

se o

f the

indi

vidu

al u

ser a

nd is

not

to b

e di

ssem

inat

ed b

road

ly.