Structural Invariance SASH-Y Running head: STRUCTURAL INVARIANCE OF THE SASH-Y
Structural Invariance of the Short Acculturation Scale for Hispanic Youth
Jeffrey A. Miller Duquesne University Andres Barona Arizona State University
Please address all correspondence concerning this article to Jeffrey A. Miller, Ph.D., Duquesne University, 106B Canevin Hall, Pittsburgh, PA 15282.
Electronic mail may be sent via
Internet to
[email protected].
1
Structural Invariance SASH-Y Abstract The Short Acculturation Scale for Hispanic Youth (SASH-Y) is a 12-item self-report measure of Hispanic acculturation for children in grades 3 through 8.
The present study explores the
psychometric qualities of the SASH-Y and provides final normative data for its use in applied and research settings. Several important research outcomes regarding the SASH-Y are elucidated.
First, the empirically demonstrated structural
invariance of the SASH-Y across two samples provides evidence for the construct validity of the scale and supports a three factor structure for the SASH-Y.
The factors include Familial
Language Use, Extrafamilial Language Use, and Ethnic Social Relations.
Second, children’s SASH-Y responses were
significantly correlated with parent report of the children’s acculturation using the Children’s Acculturation Scale (Franco, 1983).
Finally, exploratory analyses were used to provide
normative conversions of SASH-Y raw scores as well as interpretive descriptions of SASH-Y scores.
2
Structural Invariance SASH-Y Short Acculturation Scale for Hispanic Youth:
3
Psychometric
Properties and Use During Psychoeducational Evaluations Developing methods to better evaluate Hispanic youth is an important task for school psychologists.
Currently,
approximately 14% of all U.S. children are Hispanic and it is projected that by the 2020 more than one in five children in the U.S. will be of Hispanic origin (Federal Interagency Forum on Child and Family Statistics, 1997).
This article focuses on the
assessment of acculturation among Hispanic youth. Acculturation, defined as the acquisition of the values, beliefs, attitudes and behaviors of the majority culture by members of a minority culture (Garcia & Lega, 1983; Rogler, Cortes, & Malgady, 1991), is viewed as one method of improving the interpretation of psychoeducational assessment data. Legal and procedural precedents for including acculturation in the psychoeducational assessment of Hispanic youth exist. Legally, exclusionary criteria for special educational classifications require determining if cultural factors account for an observed learning problem (Individuals with Disabilities Act, 1997).
Indeed, this is one of the most enigmatic mandates
in IDEA because no specific direction is provided for determining what cultural factors are or how to determine if they account for an observed learning problem.
Procedurally,
the assessment of language minority students should also include
Structural Invariance SASH-Y an examination of acculturation.
4
The Standards for Educational
and Psychological Testing (American Educational Research Association, American Psychological Association, & American Council on Measurement, 1985) states that bilingualism is complex and involves literacy, communication, and social functions.
The standards further recommend that language
background as well as language proficiency be considered in the interpretation of test results.
Research consistently has shown
that social context and language background are factors in acculturation (Franco, 1983; Padilla, 1980; Szapocznik & Kurtines, 1993). Despite these edicts, level of acculturation does not appear to be a factor in referrals to special education although it does appear to impact on placement decisions.
In a very
revealing study, Collier (1987) examined the referral and placement decisions for 95 Hispanic elementary school students and reported no differences were noted in level of English proficiency or level of acculturation in making referral decisions.
However, referred students who subsequently were
tested and placed obtained significantly higher English proficiency and acculturation scores than those who were not placed (Collier, 1987).
These findings suggest that evaluation
teams provide services only to students who are acquiring the English language and who are acculturating, rather than
Structural Invariance SASH-Y
5
determining if language acquisition and acculturation are slowed because of a learning problem.
That is, students with primary
home languages other that English, who are less well acculturated, and who have a legitimate disability are frequently not identified. Despite the legal mandates to consider culture and research indicating the relevance of acculturation in the special education decision making process, to date a satisfactory selfreport test measuring acculturation among Hispanic children has not been developed (Negy & Woods, 1992).
Barona and Miller
(1994) specifically developed the Short Acculturation Scale for Hispanic Youth (SASH-Y) to fill this void.
Preliminary evidence
was provided for the reliability and validity of the SASH-Y and tentative norms were published (Barona & Miller, 1994).
In the
present study, additional minor changes to the wording of the SASH-Y were made.
Using the updated version of the SASH-Y, a
more thorough examination of the scale’s validity was conducted by testing its structural invariance against the Barona and Miller (1994) sample and final normative tables were developed. Method Participants Sample A: Try-out Sample.
Sample A represents the sample
of Hispanic children previously collected and the data reported by Barona and Miller (1994).
Participants included 227 public
Structural Invariance SASH-Y
6
school students in grades 3 through 8 who self-reported their ethnicity as Hispanic.
The average age of Sample A was 12.1
years (SD = 1.86, Range = 8.75 to 15.92) and 53.7% were male and 46.3% were female.
None of the participants were identified as
receiving special education, English-as-a-second-language, or bilingual education classes.
Data were collected from four
schools (2 urban, 2 suburban) in Phoenix and Mesa, Arizona. Sample B: Norming Sample. after Sample A.
Sample B was collected two years
Participants included 665 public school
students in grades 3 through 8 who self-reported their ethnicity as Hispanic.
The average age of Sample B was 11.5 years (SD =
1.80, Range = 8 to 16) and 47.4% were male and 52.6% were female.
None of the participants were identified as receiving
special education, English-as-a-second-language, or bilingual education classes.
Data were collected from nine different
schools (5 urban, 4 suburban) in Phoenix, Mesa, and Tempe, Arizona. Parent Sample.
A random sample of 300 parents of
participants in Sample B who reported that at least one of their parents spoke English was selected to respond to the Children’s Acculturation Scale (CAS; Franco, 1983) about their participating child.
Of the 300 Children’s Acculturation Scales
that were sent to parents, 149 (49.6%) were returned adequately completed and were included in the study.
Structural Invariance SASH-Y
7
Instruments Short Acculturation Scale for Hispanic Youth (SASH-Y; Barona & Miller, 1994).
The SASH-Y is a 12-item self-report
measure of Hispanic acculturation that uses a 5-point Likert scale.
Response options are given values 1 to 5 from left to
right on the likert scale.
It is a summated rating scale with
scores ranging from 12 to 60 and higher scores indicate higher acculturation to U.S. society (Barona & Miller, 1994).
Three
factors comprise the SASH-Y acculturation construct: Familial Language Use, Extrafamilial Language Use, and Ethnic Social Relations. Two modifications were made to the scale prior to its administration to the norming sample based on information gathered during data collection for the try-out sample.
First,
for items 10, 11, and 12 the word White in the response options was replaced by the word non-Hispanic.
The term non-Hispanic
was defined in the instructions to include “African American, White, American Indian, Asian American, or any other person who is not Hispanic.”
Second, the word “prefer” in item 8 was
replaced with the words “want to” and the words “you want” in item 8 were underlined.
The version of the SASH-Y reflecting
these changes and for which the normative data are presented can be seen in Appendix A.
Structural Invariance SASH-Y Children’s Acculturation Scale (CAS; Franco, 1983).
8
The
CAS is a 10-item parent report scale of Hispanic children’s acculturation.
Questions cover peer associations, ethnic
identification, and language preference, usage, and proficiency. Internal reliability was measured by means of coefficient alpha and yielded a coefficient of .77.
Franco (1983) conducted
several analyses that supported the construct and concurrent validity of the scale. Procedure Both Sample A and Sample B were collected in the same manner.
All students remained in their assigned classrooms
throughout the test administration.
The experimenters were
introduced to the students by the classroom teacher. Experimenters explained that student participation was voluntary and that students could withdraw from the experiment at any time.
Students were informed that participation in the
experiment would have no effect on their class grades and that individual responses would be kept confidential.
The
experimenters read the instructions for the instrument to each class.
Each student then completed the SASH-Y.
Experimenters
were available for questions during the completion of the instruments. A random sample of 300 parents of students who self identified as Hispanic and whose parents spoke English were
Structural Invariance SASH-Y asked to complete the CAS on their child.
9
Each of the students
was given a copy of the CAS with instructions to have one parent complete the form and return the form to the school.
Written
instructions as well as a written explanation of confidentiality were provided for the parent.
Each parent completed the
questionnaire and their child then returned the form to the school office. Results Structural Invariance Confirmatory factor analysis (CFA) using LISREL 8 (Jöreskog & Sörbom, 1996) was conducted to determine if the factor structure of the SASH-Y for Sample B was invariant with the factor structure of Sample A.
The expected factor structure
based on the factor analysis conducted by Barona and Miller (1994) of the SASH-Y was as follows:
Extrafamilial Language Use
(items 1, 4, 5, 6, 7, 8), Familial Language Use (items 1, 2, 3, 9), and Ethnic Social Relations (items 10, 11, 12). Due to the large difference in sample size between the two samples, bootstrapping was used to estimate the covariance matrix of Sample B for 220 participants.
In the bootstrap
method a random sample of size n is repeatedly drawn with replacement from the original empirical sample resulting in an estimate of the covariance matix (Bollen & Stine, 1997).
The
parameters for the bootstrap model were 200 samples of 33% of
Structural Invariance SASH-Y the total sample (N = 665, n = 220).
10
The bootstrapped sample
covariance matrix was used for subsequent structural modeling. The covariance matrix for Sample A was not positive definite; a common occurrence in data preparation for structural equation modeling with ordinal data
(Múthen, 1997).
To reduce the
impact of the positive indefinite matrix on subsequent model evaluation the unweighted least-squares method of estimation was used (Múthen, 1997). The most restrictive model of structural invariance was tested.
That is, factor loadings, factor intercorrelations, and
error variances were constrained to be equivalent.
The three
factor model of SASH-Y acculturation was determined to be invariant across the two samples as indicated by the following goodness-of-fit test results:
χ2 (128) = 89.30, p > .05,
goodness-of-fit index (GFI) = 0.99, and root mean square residual (RMR) = 0.044. Reliability and Concurrent Validity Coefficient alpha and split-half reliabilities with Spearman-Brown correction were computed.
For the norming sample
(N=665) coefficient alpha was .89 and split-half reliability was .84. Child report of acculturation on the SASH-Y was correlated with parent report of their child’s level of acculturation as measured by the CAS.
The correlation between parent and child
Structural Invariance SASH-Y
11
reports of acculturation was .46 (2-tailed significance p < .01). Normative Data Normative data were computed in two ways.
First, T-scores
and percentiles were computed for each score using a theoretical normal distribution based on the scale’s mean and standard deviation.
Second, normalized z- and T-scores were computed
based on the observed cumulative distribution of scores from the norming sample. Appendix B.
A table of normative data are presented in the
Exploratory analysis of the distribution of scores
indicated the SASH-Y has a unimodal distribution with a mean score of 40.6 and standard deviation of 10.9. measurement is .42.
Standard error of
Qualitative descriptions of score ranges
were developed based on exploratory analysis.
Respondents with
T-scores less than 40 are classified Low Acculturation. Respondents with T-scores greater than or equal to 40 and less than 55 are classified Moderate Acculturation.
Respondents with
T-scores greater than or equal to 55 are classified High Acculturation. Discussion The results of the structural invariance analysis were very strong and provide considerable evidence that the SASH-Y is a valid measure of the construct of acculturation.
Concurrent
validity evidence between parent and child reports of
Structural Invariance SASH-Y
12
acculturation was significant with a .46 correlation between parent and child ratings.
Variables that possibly attenuate
this correlation are differences between the parent and the child in perceptions, expectations of behavior, and experience with the host culture.
The normative data has been stable
across both samples and the small standard error of measurement suggests evaluators can have confidence in the accuracy of SASH-Y scores. Administration of the SASH-Y Respondents should either read or be read the instructions on the SASH-Y protocol (see Appendix A).
As well,
the items and response options may be read by either the respondent or the examiner.
Response options are given values 1
to 5 from left to right on the likert scale and the values for all 12 items are summed.
Standard scores can be found in the
norms table available in Appendix B.
It is recommended that T-
scores and percentiles be taken from the theoretical normal columns in Table B1.
The theoretical normal distribution
corrects the slight negative skew in the observed sample. Finally, determine the level of acculturation classification based on the following cut scores: T-scores less than 40 = Low Acculturation, T-scores greater than or equal to 40 and less than 55 = Moderate Acculturation, and T-scores greater than or equal to 55 = High Acculturation.
All derived scores and
Structural Invariance SASH-Y
13
classifications can then be recorded on the bottom of page two of the SASH-Y protocol. Using the SASH-Y The remainder of this article focuses on the use of the SASH-Y during psychoeducational evaluations.
It would be most
convenient if performance on the SASH-Y could be examined and based on predetermined cut scores a definitive decision regarding the extent of cultural influences on learning problems could be made.
At this point, however, such exactness does not
appear possible.
For example, a low score on the SASH-Y cannot
unequivocally rule out qualifying for special education services since it is possible for a student to obtain a low acculturation score and to qualify for special education services due to a handicapping condition.
Conversely, it is possible to rule out
cultural factors as contributing to a learning problem when a high acculturation score is obtained.
This is not always the
case because for some differential decision-making a relatively small declination in acculturation could account for a student's mild learning difficulties. These problems suggest that the best approach is to qualify only highly acculturated minority students with significant evidence of an educational handicapping condition. Unfortunately, this strategy results in what Collier (1987) observed as potentially preventing low acculturated students
Structural Invariance SASH-Y with learning difficulties from receiving needed services.
14
Too
often field based personnel have resolved this scenario by simply admitting students with a slow rate of language acquisition accompanied by other significant signs of a disability, such as a nonverbal IQ-Achievement discrepancy, into special education programs.
Such a strategy, however, does not
address the issue of whether the slow rate of cognitive-academic language proficiency acquisition is related to acculturation issues rather than related to a diagnosable disability. Further, this strategy can lead to the inappropriate exclusion of minority students from special education for several years while evaluation teams wait to determine if a student’s rate of language acquisition is indeed slow. So it seems the competent evaluator should make an idiographic decision about the relationship between culture and a student’s learning problem based on several sources of assessment data, which could include the SASH-Y.
However,
before specifying actual decision guidelines, it is important to review the relevant literature and to develop and test hypotheses about the relationship between acculturation and specific referral questions.
For example, there is considerable
evidence relating acculturation and emotional problems, in fact, Rogler (1991) argues that different emotional problems have been found to be related to varying levels of acculturation.
Structural Invariance SASH-Y
15
The literature on the relationship between level of acculturation and academic achievement has yielded mixed results.
Rodriquez (1991) found acculturation positively
related to academic achievement test performance in learning disabled Hispanic children.
Similarly, Collier (1987) found a
significant relationship between acculturation and achievement. On the other hand, Garcia-Vasquez (1995) reported that acculturation was not related to achievement scores as measured by the Iowa Tests of Basic Skills.
Rather, achievement for
Hispanic students was simply related to level of English language proficiency. Although the relationship between SASH-Y acculturation scores and specific eligibility decisions will require further exploration, the SASH-Y in it’s current form can be a significant aid in determining whether cultural factors should be considered in the interpretation of test scores.
Clearly,
students obtaining a low SASH-Y acculturation score should be evaluated by a bilingual, bicultural examiner with considerable expertise and sensitivity to the impact of cultural issues on test scores.
Evaluators who test a student with moderate
acculturation socres should seek consultation from a bilingual, bicultural psychologist and exercise caution in test interpretation.
Finally, the test scores of students with high
acculturation scores are most likely not significantly
Structural Invariance SASH-Y
16
inlfuenced by culture and can be interpreted in the usual manner. Future research in this area should include conducting an extensive review and categorizing information on the relationship between level of acculturation and educationally relevant indicators of disabilities.
Subsequently, research
should be conducted to fill the many gaps identified by the above suggested review.
The current study provides a useful
instrument for psychoeducational decision making and one that will hopefully be useful in future research to examine the many questions regarding the relationship between level of acculturation and education.
Structural Invariance SASH-Y
17
References American Educational Research Association, American Psychological Association, and American Council on Measurement. (1985).
Standards for educational and psychological testing.
Washington, DC: American Psychological Association. Barona, A., & Miller, J. A. (1994). Scale for Hispanic Youth (SASH-Y):
Short Acculturation
A preliminary report.
Hispanic Journal of Behavioral Sciences, 16, 155-162. Bollen, K. A., & Stine, R. A. (1997).
Bootstrapping
goodness-of-fit measures in structural equation modeling.
In K.
A. Bollen & J. S. Long (Eds.) Testing structural equation models.
Newberry Park, CA: Sage Publications.
Byrne, B. M. (1989).
A primer of LISREL: Basic
applications and programming for confirmatory factor analytic models.
New York:
Springer Verlag.
Collier, C. (1987). Comparison of acculturation and education characteristics of referred and non-referred culturally and linguistically different children.
Theory,
Research and Applications: Selected Papers from the Annual Meeting of the National Association of Bilingual Education, 16, 183-195. Franco, J. N. (1983). An acculturation scale fore MexicanAmerican children. The Journal of General Psychology, 108, 175181.
Structural Invariance SASH-Y
18
Garcia-Vasquez, E. (1995). Acculturation and academics: Effects of acculturation on reading achievement among MexicanAmerican students.
Bilingual Research Journal, 19, 305-315.
Garcia, M., & Lega, L. I. (1979). Development of a Cuban ethnic identity questionnaire. Hispanic Journal of Behavioral Sciences, 1, 247-261. Individuals With Disabilities Education Act Amendments of 1991, Pub. L. No. 102-119 (1994). Jöreskog, K. G., & Sörbom, D. (1996). reference guide.
LISREL 8 user’s
Chicago: Scientific Software.
Múthen, B. O. (1997). Goodness-of-fit with categorical and other nonnormal variables.
In K. A. Bollen & J. S. Long (Eds.)
Testing structural equation models.
Newberry Park, CA: Sage
Publications. Padilla, A. M. (1980). The role of cultural awareness and ethnic loyalty. In A. M. Padilla (Ed.), Acculturation: Theory, models and some new findings. Boulder, CO: Westview. Rodriquez, R. F. (1991). Sociocultural factors and achievement among Hispanic children with exceptionalities. Journal of Educational Issues of Language Minority Students, 9, 155-164.
Structural Invariance SASH-Y
19
Rogler, L. H., Cortes, D. E., & Malgady, R. G. (1991). Acculturation and mental health status among Hispanics: Convergence and new directions for research. American Psychologist, 46, 585-597. Szapocznik, J., & Kurtines, W. M. (1993). Family psychology and cultural diversity: Opportunities for theory, research, and application. American Psychologist, 48, 400-407.
Structural Invariance SASH-Y Figure Caption Figure 1.
Path diagram of LISREL estimates for the SASH-Y
structural model.
20
Structural Invariance SASH-Y
.30
.23
Item 1
Item 2 .88
.10
.30
.27
.59
Item 3 .95
.24
21
Item 4
.87 .84
Extrafamilial Language Use
.88
Item 5 .83
.30
Item 6 .88
.23
.32
Item 7
.82
.49
Item 9 .71 Item 10 .81
.35
Item 11 .78
.39
.46
Item 8 .80
.36
Familial Language Use
Item 12
Ethnic Social Relations
.48
Structural Invariance SASH-Y Appendix A
22
Structural Invariance SASH-Y
23
SASH-Y Name:____________________________ Instructions:
1.
only English E
more Spanish than English B
both Equally C
more English than Spanish D
only English E
more Spanish than English B
both Equally C
more English than Spanish D
only English E
more Spanish than English B
both Equally C
more English than Spanish D
only English E
What language(s) do you usually speak with your friends? only Spanish A
6.
English better than Spanish D
In which language(s) do you usually think? only Spanish A
5.
both Equally C
What language(s) do you usually speak at home? only Spanish A
4.
Spanish better than English B
What language(s) do your parents speak to you in? only Spanish A
3.
Circle the letter below each question that best answers the question for you. For the last three questions, non-Hispanic means African American, White, American Indian, Asian American, or any other person who is not Hispanic. Please answer every question!
What language(s) do you read and speak? only Spanish A
2.
Date:__________________
more Spanish than English B
both Equally C
more English than Spanish D
only English E
In what language(s) are the TV Programs you usually watch? only Spanish A
more Spanish than English B
both Equally C
more English than Spanish D
(please continue on page 2)
only English E
Structural Invariance SASH-Y 7.
In what language(s) are the radio programs you usually listen to? only Spanish A
8.
more Spanish than English B
both Equally C
more English than Spanish D
only English E
more Spanish than English B
both Equally C
more English than Spanish D
only English E
more Hispanic than non-Hispanic B
both Equally C
more non-Hispanic than Hispanic D
all non-Hispanic E
You like going to parties at which the people are: all Hispanic A
12.
only English E
more English than Spanish D
Your close friends are: all Hispanic A
11.
both Equally C
In what language(s) do your parents speak with their parents? only Spanish A
10.
more Spanish than English B
In what language(s) are the movies, TV, and radio programs you want to watch or listen to? only Spanish A
9.
24
more Hispanic than non-Hispanic B
both Equally C
more non-Hispanic than Hispanic D
all non-Hispanic E
more non-Hispanic than Hispanic D
all non-Hispanic E
The people you visit or who visit you are: all Hispanic A
more Hispanic than non-Hispanic B
both Equally C
For Official Use Only Raw Score:______
T-score: ______
Percentile: _____ Category:_______________
Structural Invariance SASH-Y Appendix B
25
Structural Invariance SASH-Y Table B1 Normative Data for SASH-Y
Theoretical Normal Raw 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50
T-score 24 25 26 27 27 28 29 30 31 32 33 34 35 36 37 38 38 39 40 41 42 43 44 45 46 47 48 49 49 50 51 52 53 54 55 56 57 58 59
Percentile 0 1 1 1 1 1 2 2 3 4 4 5 7 8 10 12 12 14 16 18 21 24 27 31 34 38 42 46 46 50 54 58 62 66 69 73 76 79 82
Normalized z-score -2.58 -2.41 -2.17 -1.99 -1.92 -1.74 -1.67 -1.59 -1.51 -1.44 -1.39 -1.34 -1.28 -1.23 -1.18 -1.09 -1.04 -1.00 -0.88 -0.82 -0.76 -0.69 -0.61 -0.55 -0.47 -0.38 -0.30 -0.22 -0.12 -0.06 0.00 0.07 0.14 0.19 0.33 0.46 0.63 0.80 0.99
T-score 24 26 28 30 31 33 33 34 35 36 36 37 37 38 38 39 40 40 41 42 42 43 44 45 45 46 47 48 49 49 50 51 51 52 53 55 56 58 60
Observed Percentile 0.5 0.8 1.5 2.3 2.7 4.1 4.7 5.6 6.5 7.5 8.3 9.0 10.1 11.0 12.0 13.7 15.0 16.2 18.9 20.5 22.4 24.5 27.1 29.2 31.9 35.3 38.3 41.2 45.1 47.8 49.9 52.6 55.6 57.6 63.0 67.7 73.7 78.8 83.8
26
Structural Invariance SASH-Y Table B1 (continued) 51 52 53 54 55 56 57 58 59 60
60 60 61 62 63 64 65 66 67 68
84 84 86 88 90 92 93 95 96 96
1.08 1.23 1.34 1.54 1.66 1.77 1.93 2.12 2.26 3.72
61 62 63 65 67 68 69 71 73 87
85.9 89.0 91.0 93.8 95.2 96.2 97.3 98.3 98.8 100.0
27