Вы находитесь на странице: 1из 6

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Reliability and Equivalence of the Spanish Translation of the Marital Satisfaction Inventory—Revised (MS1-R)

Charles Negy

University of Central Florida

Douglas K . Snyder Texas A&M University

This study examined the reliability and linguistic equivalence of the Spanish translation of the Marital

Satisfaction Inventory—Revised (MSI R) in a sample of 86 bilingual Mexican American couples. Overall, findings provided preliminary support for using this translation of the MSI—R with respondents whose preferred language is Spanish. Coefficients of internal consistency and temporal stability for scales constituting the Spanish MSI-R averaged .72 and .75, respectively. Linguistic equivalence coefficients for respondents completing the MSI-R in both languages averaged ,69. Mullivariate analyses of variance indicated no significant effect for language of administration on MSI-R scale scores. Discussion emphasi/cs caveats in using the Spanish MSI-R and directions for further research.

Hispanics currently constitute the second largest ethnic minority

group in the United States and are projected to become the largest

ethnic minority by 2005 (U.S. Census Bureau, 1999).' Appropriate

assessment ol" Hispanic Americans is an importan t challenge in

this age of increasing multiculturalism (Cuellar, 1998). However,

because of historical and recent trends in emigration from Latin

American countries, as many as 40% of Hispanic adults living in

the United States are either monolingual Spanish speakers or have

limited English proficiency (Garcia-Pretn, 1996). Consequently,

for mental health professionals to adequately provide services to

Spanish-dominant Hispanics, psychological measures available in

Spanish and having evidence of reliability and validity will be-

come increasingly important.

Marriage and family relationships constitute an integral facet of

many Hispanics' lives (Becerra. 1988; Falicov, 1996; Garcia-

Preto, 1996; Vega, 1990). Despite growing interest in assessment

strategies specific to couples and families (Kashy & Snyder, 1995;

Snyder, Cavell, Hcffcr, & Mangrum, 1995), few studies have

invesligated the relationships of Hispanic couples (Bean & Crane,

1996). Even fewer studies have addressed the appropriateness of

standardized marital and family assessment techniques with the

growing Hispanic population (Negy & Snyder, 1997). Although

several measures have been translated into Spanish for use with

Hispanic couples—including the Conflicl Tactics Scales (Straus,

Hamby, Boney-McCoy, & Sugarman, 1996), Dyadic Adjustment

Scale (Spanier, 1976), Family Environment Scale (Moos & Moos,

1986), and Styles of Conflict Inventory (Metz, 1993)—rarely have

adaptations of measures originally developed in English been

examined for their reliability, validity, and linguistic (alternate

This article is based on findings presented at the meeting of the Inter- American Congress of Psychology in Caracas, Venezuela, June 1999. Correspondence concerning this article should he addressed to Charles Negy, Department of Psychology, P.O. Box 161390, University of Central

Florida, Orlando, Florida 32816-1390, or to Douglas K. Snyder, Depart- ment of Psychology. Texas A&M University, College Station, Texas 77843-4235. Electronic mail may be sent to cnegy@pegasus.cc.ucf.edu or

to d-snyder@tamu.edu.

language) equivalence to their original version. The lack of nor-

mative data on Hispanic marriages precludes incorporating a

culture-specific context necessary for evaluating and intervening

with distressed couples within this ethnic group. A primary factor

limiting both the clinical and research literature has been the

failure to examine standardized measures of marital functioning

with Hispanic American couples (Knight, Tein, & Shell, 1992;

Vega, 1990).

This study examined the reliability and linguistic equivalence of

the Spanish translation of the Marital Satisfaction Inventory—

Revised (MSI-R; Snyder. 1997). The MSI-R is a 150-item, true-

false, self-report measure of relationship functioning designed to

identify both the nature and intensity of distress in distinct areas of

partners' interaction. The MSI—R includes 2 validity scales, 1

global distress scale, and 10 additional scales assessing specific

dimensions of the relationship. Previous studies have supported the

internal consistency and temporal stability of MSI-R scales, as

well as their convergent and discriminant validity (see Snyder &

Aikrnan, 1999, for a review). Specifically, coefficients of internal

consistency for the MSI-R scales derived from a combined sample

of 2,040 persons from the community and 100 persons in couple

therapy ranged from .70 to .93 (M = .82); temporal stability

coefficients derived from 210 persons from the community re-

.74 to .88 (M = .79). In a

from

tested

comparison of 50 clinic couples and 77 community couples

matched on demographic indexes, each of the MSI-R scales dis-

criminated between the community and clinic couples at/? < .001,

with moderate-to-large effect sizes (Cohen's d) ranging from 0.43

to 2.35 (M = 1.07). Actuarial tables linking scale scores to

descriptors of the relationship provided by clinicians and both

spouses showed the MSI-R scales to relate to a broad range of

after

6 weeks

ranged

1 The term Hispanic is not preferred by all members of this ethnic group. However, to eliminate potential confusion by using other terms inter-

changeably (e.g

when referring to persons who trace their heritage to Spanish origins. The exception to this is persons of Mexican descent living in the United States,

to whom we refer as Mexican Americans.

Latino), Hispanic will be used throughout this article

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

426

NfiGY AND SNYDER

external criteria consistent with their interpretive inlenl (Snyder,

1997).

A modest body of literature has emerged examining the useful-

ness of the MSI-R in cross-cultural applications. Klann, Hahlweg,

and Hank (1992) developed a German adaptation of the original

MSI and garnered evidence for both the reliability and validity of

this translation in community and clinical settings. A more recent

analysis of their data scored on the revised MSI-R scales yielded

internal consistency coefficients rangin g from .74 to .92 (M = .83)

and confirmed these scales' ability to discriminate between Ger-

man community and clinic couples with effect sizes ranging

from 0.26 to 1.22 (M = 0.75; Abbott Snyder, Cleaves,

& Klann, 2000). Spasojevic, Heffer, and Snyder (2000) examined

a Bosnian translation of the MSI-R for use with U.S. refugee

couples and documented the linkage of relationship difficulties to

posttraumatic stress in this population. Measures of affeclivc and

problem-solving communication, quality of leisure time together,

disagreement about finances, and dissatisfaction with the sexual

relationship con-elated significantly (p < .01) with measures of

posttraumatic stress disorder (PTSD) symptomatology for both

husband s and wive s (>s rangin g fro m .42 to .74). In a stud y

comparing 75 Mexican American and 66 non-Hispanic White

American couples on the English version of the MSI-R , Negy and

Snyder (1997) found preliminary evidence for the structural equiv-

alence of MSI-R scales across both groups; moreover, multivariate

analyses indicated the absence of a significant effect for ethnicity

on mean scale profiles. In a follow-up study comparing these

monoethnic couples with 72 interethnic couples in which one

partner was Mexican American and the other was non-Hispanic

White, Negy and Snyder (2000) determined that the interethnic

couples were more similar to non-Hispanic White couples than to

Mexican American couples across multiple domains assessed by

the MSI-R.

studies examining psychometric

characteristics of the Spanish translation of the MSI-R (Snyder,

1996). Such findings are critical not only for use of the MSI-R

with Hispanics in the United States for whom Spanish is the

preferred language but also for studies exploring cross-cultural

Hahlweg ,

Lacking

thus

far

have

been

differences and similarilics in couples 1 relationships. Concern

about using instruments conceived and validated in F.nglish wilh

individuals having limited English proficiency has been empha-

sized repeatedly in the literature (Bracken & Barona, 1991; Cuel-

lar, 1998; Malgady , Rogler, & Costantino, 1987; Okazaki & Sue,

(995). Studies of internal consistency, temporal stability, and

linguistic equivalence constitute a critical initial phase of the test

adaptation process (Butcher & Han, 1996; Geisinger, 1994).

Method

Sample

The sample included 86 bilingual Mexican American couples recruited from a community in the southwest United States in which Mexican

Americans constitute

students in psychology recruited couples from the local community and

were prohibited from recruiting members

but were otherwise free to draw on their own personal and organizational

contacts in the community. Couples received no remuneration for their participation. Given the frequent reservation of minority group members to participate in research (of, Oka/aki & Sue, 1995). ii was anticipated that this recruitment strategy might facilitate rather than detract from sample

family

population. 2 Graduate

approximately

85%

of

the

of their own immediate

representativeness. Participants had to be at least IN years of age, legally married, and able to read and speak both English and Spanish. The use of bilingual samples has been noted as a preferred method for examining the

comparability of translated tests (Bracken & Barona, 1991).

Husbands and wives averaged 38 and 36 years in age, respectively, and had been married an average of 12 years. Most respondents (72%) reported having one or more children. The majority of participants had graduated

from high school and had some college education (M = 14.3 years of education). The majority of husbands (75%) and wives (69%) were em- ployed outside the home. The sample varied broadly in occupational status but had a tendency toward higher level positions (59% professional, 13%

administrative, 19% clerical or technical, and 9% manual labor). On none

of these variables were there significant gender differences (all /?s >

.15).

Measures

Marital Satisfaction

InventoryRevised

All respondents completed the MSI-R (Snyder. 1997) either (a) twice in cither the original English or Spanish translation or (b) once each in both languages. Scale names, abbreviations, and brief descriptions are as follows. Inconsistency (INC). A validity scale assessing the individual's con- sistency in responding to item content. (High scores reflect greater

inconsistency.) 3

A validity scale assessing individuals'

Conventionalization

(CNV).

tendencies to distort the appraisal of their relationship in a socially desir- able direction. (High scores reflect denial of common relationship shortcomings.) Global Distress (CDS). Measures individuals" overall dissatisfaction

with the relationship. Affective Communication (AFC]. Evaluates individuals'

dissatisfaction

with the amount of affection and understanding expressed by their partner.

Problem-Solving Communication (PSC). Assesses the couple's gen- eral ineffectiveness in resolving differences. Aggression (AGG). Measures the level of intimidation and physical aggression experienced by respondents from their partner. 'Time 'together (TTO). Evaluates the eouple's companionship as ex- pressed in time shared in leisure activity.

Disagreement About Finances (FIN). Measures relationship discord regarding the management of finances. Sexual Dissatisfaction (SEX). Assesses dissatisfaction with the fre-

quency and quality of intercourse and other sexual activity. Role Orientation (ROR). Evaluates the respondent's advocacy for a traditional versus nontraditional orientation toward marital and parental

gender roles. (High scores reflect a nontraditional orientation.)

Family History of Distress (FAM). Reflects the disruption of relation- ships within the respondent's family of origin.

Assesses the relationship quality

between respondents' and their children as well as parental concern regard- ing one or more child's emotional and behavioral well-being. Conflict Over Child Hearing (CCR). Evaluates the extent of conflict between partners regarding child rearing practices. The MSI-R is administered to each partner separately and requires approximately 25 min to complete. Individuals' responses are scored along

Dissatisfaction With Children (DSC).

wilh samples de-

scribed

^ The INC scale comprises 20 pairs of items in which the two items constituting each pair are either (a) similar in content and should be answered in the same direction or (b) dissimilar or nearly opposite in

content and should be answered in opposite directions. Item pairs were sampled across the entire range of construct domains reflected in (he MST—R item pool. Hence, coefficients of internal consistency are not appropriate measures for this scale.

None of the participants in this study overlapped

in Negy and Snyder (1997) or Negy and Snyder (2000).

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

MSI-R SPANISH

VERSION

427

the 13 profile scales and are plotted on a standard profile sheet using gender-specific norms. Each of the scales, excluding Ihc validity scales (INC and CNV) and ROR, are scored in a direction whereby higher scores

reflect higher levels of relationship distress.

Acculturation Rating Scale for

Mexican-Americans

Respondents also completed the Acculturation Raring Scale for

Mexican-Americans (ARSMA; Cuellar. Harris, & Jasso, 1980). The

a scale ranging from 1 to 5; the

total score indicates the mean rating across items. High or moderately high

scores (>3.2) reflect relatively high levels of acculturation to predomi- nantly White American culture, whereas low or moderately low scores «2.8 ) indicate low levels of acculturation. Factor analysis of the ARSMA

has identified four dimensions underlying the scale; (a) Language Famil- iarity, Usage, and Preference; (b) Ethnic Identity and Generation; (c) Reading, Writing, and Cultural Exposure; and (d) F.thnic Interaction. The ARSMA has been shown to have high internal consistency (a — .88) and

good stability over a 1-month interval (r = .80). Various studies have supported the scale's validity as a measure of acculturation among Mexi- can Americans (Cuellar, Arnold, & Maldonado, 1995; Rogfer, Cortes, & Malgady, 1991). Participants in this study showed a broad range of accul- turation levels, with scores on the ARSMA ranging from 1.30 to 4.05

(M = 3.00, S D - 0.56).

ARSMA contains 20 items, each scored on

Procedure

The Spanish translation of the MSI-R was developed by a team of three bilingual psychologists with previous experience in translating psycholog-

ical measures from English to Spanish. Back translations and further

revisions were conducted until consensus was reached on the linguistic equivalence of items. The Spanish version of the MSI-R is intended for use

with Spanish-speaking individuals from diverse regions and has been reviewed by psychologists working with couples in the southwestern United States, Mexico, and Spain.

Participants completed the MSI-R on two different occasions with a 7-10 day interval between Time 1 and Time 2. Interviewers provided the study rationale, obtained informed consent, and administered all measures to couples in their homes. Spouses completed measures separately and without collaboration. Participants were randomly assigned to one of four

groups reflecting language and order of presentation: (a) Spanish-English (21 couples), {b) English-Spanish (23 couples), (c) Spanish-Spanish (23 couples), and (d) English-English (21 couples). Scores from the first two groups were used to derive linguistic equivalence (alternate-language re-

liability) Pearson product moment correlation coefficients: these were compared with (a) within-language test-retest reliability coefficients de- rived from the second two groups and (,b) test—relest reliability data from the standardization sample (Snyder, 1997). Responses to items from all

four groups at Time 1 were used to compare estimates of internal consis- tency (Cronbaeh's alpha) for the Spanish and English administrations in this study and to compare these to internal-consistency coefficients derived from the standardization sample. Finally, scores from Time 1 for all four groups were used to compare effects of language and order of presentation

on mean scale scores.

Results

Alpha coefficients of internal consistency for the study samples

and the MSI-R standardization sample are shown in Table 1.

Alpha coefficients for the Spanish MSI-R ranged from

(M = .72), and those for the English MSI-R ranged from .72 to .90

(M = .82), compared with a range of .70 to .93 (M = .82) for the

standardization sample on the English version. For the Spanish

MSI-R, all alpha coefficients were above .60 with the exceplion of

that for the DSC scale (a = .22). However, by using Fisher's

.22 to .89

Table 1

Internal-Consistency Coefficients

of MSI-R

Scales in Study

Samples and MSI-R

Standardization

Sample

Scale

Spanish' 1

English* 1

Standard''

Inconsistency (INC) Conventionalization (CNV)

.80

.88

.83

Global Distress (GDS)

.89

.91)

.93

Affective Communication (AFC)

.83

.86

.85

Problem- Solving Communication (PSC)

.86

.89

.89

Aggression (AGO)

.79

.88

.81

Time Together (TTO)

.68

.80

.SO

Disagreement About Finances (FIN)

.68

.72

.79

Sexual Dissatisfaction (SEX)

.80

.79

.84

Role Orientation (ROR)

.73

.75

.83

Family History of Distress (FAM)

.75

.78

.78

Dissatisfaction With Children (DSC)

^ 2

.73

.70

Conflict Over Child Rearing (CCR)

.61

.83

.78

Nole. MSI-R = Marital Satisfaction Inventory—Revised; Standard — standardization sample. Dashes indicate that internal consistency does not apply to the Inconsistency scale. a n - 88 persons (66 for DSC and CCR). b n - 84 persons (57 for DSC and CCR). c n =2.14 0 persons (930 for DSC and CCR).

r-to-Z transformations to compare alpha coefficients across the

two language versions, we found lower coefficients for the Spanish

MSI-R on the AG O scale (p < .05) and on the tw o child-related

scales: DSC and CCR (ps < .01).

Additional analyses were conducted to examine potential effects

of gender, acculturation, and time of administration on alpha

coefficients. Compared with alpha coefficients derived from the 44

couples completing the Spanish MSI—R at Tim e I , alpha coeffi-

cients derived from the 44 couples completing the Spanish MSI-R

at Time 2 (including 23 couples completing the Spanish version on

both occasions) were higher for the DSC scale (a = .57

.22,

vs.

p ~ .01). Among the 44 couples completing the Spanish MSI-R at

Time I, husbands obtained a higher alpha coefficient than did

.65, p = .01) but a lower

alpha coefficient on the DSC scale (re = —.10

Among these same 44 couples, by using a median split to divide

participants into low- and high-acculturation groups (ARSM A

scores of 3.03 and 3.10 for husband s and wives, respectively), w e

found that individuals in the low-acculturation group obtained a

higher alpha coefficient than respondents in the high-acculturation

group on AFC (a = .88

.40, p = .04). Noteworthy

among these effects was the consistently lower alpha coefficient

for DSC among men, individuals lower in acculturation, and

participants completing the MSI-R for the first time when appre-

hension might be higher.

Tesl-tetest reliability coefficients of temporal stability for the

study samples and the MSI-R standardization sample are shown in

Table 2. Tcsl-retest coefficients for the Spanish MSI-R ranged

from .37 to .92 (M — .75). and those for the English MSI-R ranged

from .57 to .91 (M = .82), compared with a range of .52 to .88

(M = .77) for the standardization sample on the English version.

Again, with the exception of lower temporal stability coefficients

(ps < .05) on the tw o child-related scales for the Spanish version,

these reliability coefficients were generally comparable across the

three groups.

Additional analyses were conducted to examine potential effects

of gender and acculturation on test-retest reliability of the Spanish

coefficient on DS C (a = -.04

.75, p = .03) but a lower alpha

wives on the SE X scale (a = .85

.04).

vs.

vs.

.41, p

=

vs.

vs.

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

428

NEGY AND

Table 2 Temporal-Stability Coefficients

Samples and MSl-R Standardization

of MSI-R

Scales in Studv

Sample

Scale

Spanish 11

English'

Standard'

Inconsistency

(INC)

.69

.57

,52

Conventionalization (CNV)

.86

.91

.78

Global Distress (CDS)

.75

.78

.74

Affective Communication (AFC)

.86

.83

.79

Problem-Solving Communication (PSC)

.92

.91

.82

Aggression (AGG)

.88

.87

.81

Time Together (TTO)

.79

.78

.77

Disagreement About Finances (FIN)

.75

.87

.74

Sexual Dissatisfaction (SEX)

.86

.85

.81

Role Orientation (ROR)

.73

.78

.88

Family History of Distress (FAM)

.76

.82

.84

Dissatisfaction With Children (DSC)

.37

.80

.79

Conflict Over Child Rearing (CCR)

.50

.83

.74

Note.

= standardization sample. " n = 46 persons (34 for DSC and CCR). =

and CCR).

MST-R

•• n

Marital

Satisfaction

Inventory—Revised; Standard

~

" n = 42 persons (31 for DSC

and

CCR).

210 persons (153 for DSC

MSl-R. Compared with their wives, husbands obtained lower

.02).

Individuals in the low-acculturation group obtained lower test—

retesl coefficients on INC (r = .43

- Linguistic equivalence coefficients (alternate-form reliability) between the English and Spanish versions of the MSl-R are shown in Table 3. Pooling the English-Spanish and Spanish-English subgroups across order of presentation, we found equivalence

coefficients ranging from .38 to .79 (M — .69) across scales; with the exception of the INC scale, all equivalence coefficients were above .60. Equivalence coefficients across scales tended to be somewhat lower for participant receiving the Spanish translation

.73), although no significant differences in

coefficients emerged for order of presentation. The mean shared variance across MSI-R scales for test-retest comparisons across language (48%) reflected a modest decline from the mean shared

variance across scales for test-retest coefficients within the Span- ish version (56%) and English version (59% and 67% for the standardization and English—English study samples, respectively). Again, additional analyses were conducted to examine potential effects of gender and acculturation on linguistic equivalence co- efficients for participants completing the MSI-R in both lan- guages. Compared with their wives, husbands obtained lower

first (mean r of .64 vs.

(r

temporal stability coefficients on CCR (r = .16

vs.

=

.79, p

=

vs.

.89, p

.001) and CCR

.14

vs.

.82, p = .004).

linguistic equivalence coefficients on CNV (r = .63

vs.

.85. p —

.01), AFC (r =

.62

vs.

.82, p = .03). ROR (r

- .64

vs.

.83. p =

.03), and CCR

(r = .15

vs.

.67, p = .01). Individuals in the

low-acculturation group obtained lower linguistic equivalence co-

efficients on FIN (r = .53

.82, p = .03). Finally, we used rnultivariate analyses of variance (MANOVAs) to compare MSI-R mean scale scores across the four language- order subgroups shown in Table 4. Using scores from Time 1 and collapsing across groups receiving either the English version or Spanish translation, a MANOVA of the 13 MSI-R scales demon- strated a nonsignificant main effect for language of administration, Wilks's lambda f"(13, 109) = 1.14,p = .34, R- = .01. Follow-up univariate comparisons confirmed the absence of significant mean

.85, p = .002) and DSC (r = .56 vs.

vs.

SNYDER

differences and small effect sizes between the English and Spanish versions for each MSI-R scale (all ps > .05, all K 2 s < .03). Additional results indicated a significant but small main effect for time of administration, Wilks's lambda r v (13. 107) — 3.73, p < .000, R 2 = .03, but no significant interaction between language and order of administration, Wilks's lambda F(39, 317) — 1.28, p — .128, R 2 — .01. Farther univariate analyses examining the

effect for time revealed that independent of language version, on the first MSI-R administration, participants showed somewhat greater inconsistency, lower defensiveness on CNV. and higher

on PSC, AGG, and SEX (all ps < .05, « 2 s

ranging from .03 to .08). In all instances of significant effects across time, mean score differences were modest: ranging from 1

levels of distress

to 3 7"-score points (mean difference — 1.65 points).

In contrast to their effects on internal consistency, temporal stability, and linguistic equivalence, neither gender nor level of acculturation had significant effects on mean MSI-R scale scores. The MANOVA for gender was nonsignificant, Wilks's lambda

F(13, 109) = 0.71, p ~ .75, R 2 = .01. with follow-up univariate comparisons confirming no significant effects for any scale (all ps > .05, all f? 2 s < .03). Similarly, the MANOVA for accullura-

tion level was nonsignificant, Wilks's lambda F(I3 , 109} — 1.75,

p — .06, /? 2 ~ .01, with follow-up univariate comparisons con-

firming no significant effects for any scale (allp s > .05, all R 2 s < .03. However, similar to effects of low education for ethnic mi- norities in the MSI-R standardizationsample (Snyder, 1997), there was a small but significant effect for participants in this study with 12 or fewer years of education to be less consistent in their responding, espouse more traditional role orientations, and report higher aggression and more discord in their family of origin,

all ps < .05, R 2 s ranging from .04 to .07. for the MANOVA based

on 13 scales, Wilks's lambda F(13. 109) = 1.82, p = .05, R 2 = .03.

Discussion

The present study using bilingual couples offers further support for using the MSI-R with Hispanic couples and provides prelim-

Table 3 Linguistic Equivalence English and Spanish

of MSl-R Scales Across Versions

 

Spanish-

English-

Scale

English' 1

Spanish 11

Combined 0

Inconsistency (INC)

.22

.51

Conventionalization (CNV)

.72

.82

Global Distress (GDS)

.54

.74

Affeclivc Communication (AFC)

.72

.72

Problem-Solving Communication (PSC)

.85

.74

Aggression (AGG)

.61

.82

Tim e Together (TTO)

.67

.81

Disagreement About Finances (FIN)

.68

.66

Sexual Dissatisfaction (SEX)

.73

.71

Role Orientation

(ROR)

.72

.84

Family History of Distress

(FAM)

.70

.81

Dissatisfaction

Widi Children

(DSC)

.71

.65

Conflict Over Child Rearing

(CCR)

.39

.72

.38

.78

.65

.70

.79

.75

.72

.67

.72

77

.75

.68

.60

Note. MSI-R - Marital Satisfaction Inventory—Revised.

" n = 42 persons (34 for DSC and CCR). h n = 42 persons (26 for DSC

and CCR). '" n = 84 persons (58

for DSC and

CCR).

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

MSI-R SPANISH VERSION

Table 4

MSI-R

Mean Scale Score Comparisons Across Four Language

X Order Study Samples

429

 

Spanish-English^

English-Spanish 11

Spanish-Spanish'"'

English-English 11

Scale

M

SD

M

SD

M

SD

M

SD

Inconsistency (INC)

54.31

9.22

52.29

9.66

52.57

9.73

50.17

9.56

Conventionalization

(CNV)

53.52

9.42

51.48

10.76

54.46

9.16

52.83

10.67

Global Distress(GDS) Affective Communication

49.24

8.76

51.21

8.36

51.09

8.08

49.71

7.82

(AFC)

49.55

9.90

48.10

10.59

49.93

8.47

47.79

8.41

Problem-Solving Communication (PSC)

48.05

9.76

50.93

10.00

48.54

8.42

47.48

9.54

Aggression (AGO)

51.60

7.81

52.31

10.74

50.89

9.93

50.38

10.41

Time Together (TTO) Disagreement About

51.50

7.18

50.02

9.47

48.74

8.54

48.36

9.64

Finances (FIN)

49.88

7.93

50.33

9.78

49.33

8.71

50.07

7.43

Sexual Dissatisfaction

(SEX)

47.33

8.28

46.79

8.74

48.26

8.15

47.05

7.75

Role Orientation (ROR) Family History of

51.26

7.12

53.36

7.92

51.76

7.20

54.12

7.06

Distress (FAM )

48.33

7.87

47.07

9.40

48.22

9.01

48.83

9.15

Dissatisfaction With Children (DSC) Conflict Over Child

50.22

7.83

51.85

10.00

48.76

7.00

49.58

9.65

Rearing(CCR)

51.53

7.92

53.19

10.38

49.50

6.36

49.23

9.57

Note. Table entries reflect scores from the first administration. " n = 42 persons (32 for DSC and CCR). b n = 42 persons (26 for DSC and CCR). DSC and CCR). d n = 42 persons (31 for DSC and CCR).

i

— 46 persons (34

for

inary evidence regarding the psychometric equivalence of the

Spanish adaptation of this measure. Most of the MSI-R scales in

the Spanish translation obtained indices of internal consistency and

temporal stability comparable to their English counterparts in both

the present study samples and the MSJ-R standardization sample.

Moreover, linguistic equivalence indices demonstrated only mod-

est decline when comparing alternate-language coefficients to

within-language coefficients derived from either repeated English

or Spanish administrations. Perhaps most important, both multi-

variate and univariate analyses confirmed the absence of signifi-

cant mean score differences between the English and Spanish

versions on any MSI-R scale.

An important exception to these overall findings involved

poorer internal consistency and temporal stability for the Spanish

translation of the two child-related scales: DSC and CCR. Reex-

ami nation of items constituting these two scales failed to identify

any clear shortcomings in the translation process from a linguistic

perspective, suggesting that findings for these scales may reflect

differences related to child rearing among Hispanic couples. In

their study of Mexican American couples using the English ver-

also oblaincd slightly

lower alpha coefficients for DSC and CCR (.63 and .68, respec-

sion of the MSI-R , Negy and Snyder (1997)

tively), concluding that interpretation of these scales for Hispanic

couples "should be undertaken with caution" (p. 419). Although

internal-consistency coefficients for the English version of DSC

and CCR with Hispanic couples were higher in the present study

(.73 and .88) and temporal stability coefficients for the English

version of these scales also were good (.80 and .83), the weaker

internal-consistency and temporal reliability findings for the Span-

ish translation of CCR and especially DSC emphasize the need for

continued caution.

The influences of both gender and level of acculturation on

psychometric properties of DSC and CCR bear further investiga-

tion. In the present study, half of all gender or acculturation effects

on internal consistency, temporal stability, and linguistic equiva-

lence were obtained for one of these two child-related scales.

Specifically, the meaning of scores on DSC and CCR derived from

the Spanish MSI-R remains uncertain for husbands and for indi-

viduals low in acculturation.

Although multivariale analyses demonstrated an effect for time

of administration regardless of language order, these effects were

restricted to a subset of MSI-R scales, and Ihe magnitude of the

time effect was modest, averaging less than 2 T-score points.

Moreover, these results are consistent wilh previous studies indi-

cating somewhat higher scores on the initial administration of

psychological measures in community samples (Finch, Savior,

Edwards, & Mclntosh, 1987; Michael & Merrell, 1998; Negy,

Lachar, & Grubcr. 1998; Reynolds & Graves, 1989).

The absence of a significant effect for language of administra-

tion on mean scale scores observed in this study, along with the

comparability of mean profiles between Mexican American and

Anglo American couples observed by Negy and Snyder (1997) for

the English MSI-R, provides preliminary support for the equiva-

lence of MSI-R scales when adapted for use with Spanish-

dominant couples. However, evidence for the scalar equivalence

(Butcher & Han, 1996) of the Spanish translation of the MSI-R

will require criterion-validation studies that link MSI-R scale

scores from Hispanic respondents to independent ralings of rela-

tionship satisfaction similar to those used in validating the MSI-R

with predominantly non-Hispanic White American couples. Find-

ings from such studies arc critical to linking a given deviation from

a scale's mean to a given level of distress in that domain.

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

430

NEGY AND SNYDER

The clinical use of instruments adapted for ethnic minorities or

cross-cultural application requires sensitivity to cultural differ-

ences as well as determination of the respondent's preferred lan-

guage and degree of acculturation. Although efforts were made in

the present translation to use standard Spanish that would be

understandable to most Spanish speakers, given the vast geo-

graphic distribution of Spanish-dominant cultures internationally

as well as ethnic differences among Spanish-speaking groups

within the United States, some degree of vernacular idiosyncracies

influencing response to the Spanish translation of the MSI-R

might be expected. Additional findings bearing on cross-cultural

utility can be anticipated from research currently under way using

the MSI-R Spanish version with both community and clinic cou-

ples in Spain and with community couples in Mexico.

References

Abbott. B. V., Snyder, D. K., Cleaves, D. H,, Hahlweg, K., & Klann, N. (2000, August). Crow-cultural reliability and validity of the German Marital Satisfaction InventoryRevised. Paper presented at the 108th

Annual Convention of the American Psychological Association, Wash- ington, DC.

Bean. F. D., & Crane, D. R. (1996). Marriage and family therapy research with ethnic minorities: Current status. American Journal of Family Therapy. 24, 3-8. Becerra. R. M. (1988). The Mexican American family. In C. H. Mindel, R. W. Habcnstcin, & R. Wright Jr. (Eds.), Ethnic families in America:

Patterns and variations (3rd ed., pp. 1141-1159). New York; Elsevier. Bracken, B. A., & Barona, A. (1991). State of the art procedures for translating, validating and using psyc coeducational tests in cross- cultural assessment. School Psychology International, 12, 119-132.

Butcher, J. X

& Han, K. (1996). Methods of establishing cross-cultural

equivalence. In J. N. Butcher (Ed.). International adaptations of the MMPI—2: Research and clinical applications (pp. 44 63). Minneapolis:

University of Minnesota Press.

Cuellar, I. (1998). Cross-cultural clinical psychological assessment of Hispanic Americans. Journal of Personality Assessment, 70, 71-86. Cuellar, I., 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.

Harris, L. C., & Jasso. R. (1980). An acculturation scale for

Mexican American normal and clinical populations. Hispanic Journal of Behavioral Sciences, 2, 199-217,

Cucllar, I

Falicov, C. J. (1996). Mexican families. In M. McGoldrick, J. Giordano, &

J. K. Pearce (F,ds.), Ethnicity & family therapy (2nd ed,, pp. 169-182).

New York: Guilford Press.

, Children's Depression Inventory: Reliability over repeated administra- tions. Journal of Clinical Child Psychology, 16, 339-341. Garcia-Preto, N. (1996). Latino families: An overview. TnM. McGoldrick,

Finch, A. J., Saylor, C. F., Edwards, G. 1

(1987).

& Mclntosh, J. A.

J. Giordano. & J. K. Pearce (Eds.), Ethnicit\ & family therapy (2nd ed.,

pp. 141-154). New York: Guilford

Geisinger, K. F. (1994). Cross-cultural normative assessment: Translation

and adaptation issues influencing the normative interpretation of assess-

Press.

ment instruments. Psychological

Assessment, 6, 304-312.

Kashy, D. A., & Snyder, D. K. (1995). Measurement and data analytic issues in couples' research. Psychological Assessment, 7, 338-348.

Hahlweg, K., & Hank. G. (1992). Deutsche validierung des

"Marital Satisfaction Inventory" (MSI) von Snyder (1981) [German validation of the "Marital Satisfaction Inventory" (MSI) of Snyder (1981)). System Familie, 5, 10-21.

& Shell. R. (1992). The cross-ethnic equivalence

Klann, N

Knight. G. P

Tein. J. Y

of parenting and family interaction measures among Hispanic and Anglo-American families. Child Development. 63, 1392-1403.

Malgady, R. G., Rogler, L. H., & Costantino, G. (1987). Ethnocultural and

linguistic bias in mental health evaluations of Hispanics. American Psychologist, 42, 228-234.

personal rela-

tionships. Palo Alto, CA: Consulting Psychologists Press.

Metz, M. E. (1993). Styles of Conflict

Inventory

(SCI) for

Michael, K. D., & Merrell, K. W. (1998). Reliability of children's self-

reported internalizing symptoms over short- to medium-length time

intervals. Journal of the American Academy of Child and Adolescent

Psychiatry.

37,

194

201.

Moos, R. H., & Moos. B. S. (1986). Family

(2nd ed.). Palo Alto. CA: Consulting Psychologists Pres,s. Ncgy, C., Lachar. D., & Gruber, C. P. (1998). The Personality Inventory

Environment Scale manual

for Youth (PIY)—Spanish version: Reliability and equivalence to the

English version. Hispanic Journal of Behavioral Sciences, 20, 391-404.

Negy, C., & Snyder. D. K. (1997). Ethnicity and acculturation: Assessing

Mexican American couples' relationships using the Marital Satisfaction

Inventory—Revised. Psychological Assessment, 9, 414-421.

Xegy, C., & Snyder. D. K. (2000'). Relationship satisfaction of Mexican-

and non-Hispanic While-American intcrethnic couples: Issues of accul-

turation and clinical intervention. Journal oj Marital and Family Ther-

apy, 26, 293-304,

Okazaki, S

& Sue, S. (1995). Methodological issues in assessment re-

search with ethnic minorities. Psychological Assessment, 7, 367-375.

Reynolds, W. M., & Graves, A. (1989). Reliability of children's report of

depressive symptomatology. Journal of Abnormal Child Psychiatry, 17,

647-655.

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

mental health status among Hispanics. American Psychologist, 46, 585-

597.

Snyder, D. K. (1996), Inventario de Satisfaction Marital, Rcvisado (MSI-

Rogler, L

R). Los Angeles. CA: Western Psychological Services.

Snyder, D, K. (1997). Manual for the Marital Satisfaction Inventory

Revised. Los Angeles, CA: Western Psychological Services.

Snyder, D. K., & Aikman, G. A. (1999). The Marital Satisfaction Inven-

tory—Revised. In M. E. Maruish (Ed,), Use of psychological testing for

treatment planning and outcomes assessment (2nd ed

pp. 1173 1210).

Mahwah, NJ: Erlbaiim.

Snyder, D. K., Cavell, T. A., Heffer, R. W., & Mangrum, L. F. (1995).

Marital and family assessment: A mukifaceted, multilevel approach. In

R. H. Mikesell, D. D. Lusterman, & S. H. McDaniel (Eds,), Integrating

family therapy: Handbook offamilv psychology and systems theory (pp.

163-182). Washington, DC: American Psychological Association.

Spanier, G. B. (1976). Measuring dyadic adjustment: New scales for

assessing the quality of marriage and similar dyads. Journal of Marriage,

and the Family, 38, 15-28.

Spasojevic, J., Heffer, R. W., & Snyder. D. K, (2000), Effects of posttrau-

matic stress and acculturation on marital functioning in Bosnian refugee

couples. Journal of Traumatic Stress, 13, 205-217.

Hamby, S. L., Boney-McCoy, S,, & Sugarman, D, B. (1996).

The Revised Conflict Tactics Scales (CTS2): Development and prelim-

inary psychometric data. Journal of Family Issues, 17, 283-316.

Straus. M. A

U.S. Census Bureau. (1999, December). Population projections of the

United Stares by age, sex, race, Hispanic origin, and nativity; 1999 to

2100. Summary Table NP-T4-B [on-line]. Washington, DC. Available

on the World Wide Web: http://www.census.gov/population/projcclions/

nation/summaryAip-t4-b.pdf.

Vega, W. A. (1990). Hispanic families in the 1980s: A decade of research.

Journal of Marriage and the Famil\. 52. 1015-1024.

Received March 10, 2000

Revision received July 24, 2000

Accepted July 25, 2000