The responses of Mexican-American children to cognitive demands in a testing situation

Date

1979

Journal Title

Journal ISSN

Volume Title

Publisher

Abstract

The purpose of this study was to examine children's behaviors in an IQ testing situation from three interrelated perspectives. One perspective examined the impact of an educational intervention program on children's test-taking behaviors. The second perspective examined the impact of age on children's test-taking behaviors. The third perspective compared the extent to which test-taking behaviors and IQ scores predicted program condition. This study followed the coding procedures developed by Hertzig, Birch, Thomas, and Mendez (1968) who investigated behavioral style differences between Puerto Rican workingclass and Anglo middle-class children in a test-taking situation. The sample for this study consisted of 40 low-income Mexican- American children, ages 4 and 5. These children participated in the treatment or control condition of the Parent Child Development Center (PCDC), University of Houston from the time they were 1 year of age, when they were randomly assigned to either condition. The instruments used in this study were the Wechsler Preschool and Primary Scale (WPPSI) and a behavioral coding scheme developed by Hertzig et al., (1968). Detailed protocols of the children's behaviors during the administration of the WPPSI were recorded by an independent observer and then coded. The coding scheme categorized responses as work or network responses; these could be expressed verbally or through action. Five research hypotheses were posed. The first hypothesis examined specific work and not-work behaviors exhibited by the children in the two groups. This hypothesis was tested using MANOVA. The results of this analysis indicated overall significant differences in test-taking behaviors between the treatment and control condition. The univariate I? tests indicated treatment-condition children exhibited more total work and were more verbal and less substitutive than control-condition children. The second hypothesis examined specific work and not-work behaviors in a developmental context through MANOVA. The results of this analysis indicated significant differences between the two age groups. The univariate F tests indicated the older children exhibited more total work and were more verbal and less substitutive than the younger children. The third hypothesis tested for an interaction between age and program condition through MANOVA. The overall F test indicated nonsignificant differences for this interaction. In other words, program impact remained constant on the children’s test-taking behaviors for both age groups. The fourth hypothesis examined children’s verbal spontaneous extensions in relation to group condition using chi-square. The results of this analysis indicated that the number of treatment-condition children who made at least one verbal spontaneous extension was significantly higher than those in the control condition. The fifth hypothesis focused on the extent to which two classes of variables (i.e., verbal and performance IQ versus test-taking behaviors) predict program condition individually and over and above their separate effects. Two models of stepwise regression were utilized to assess their contribution to the criterion variable. The results indicated that both classes of variables contributed a significant portion of the variance when tested individually. However, test-taking behaviors also explained a significant portion of the variance in program condition over and above that accounted for by IQ scores. On the other hand, IQ scores failed to contribute significantly to the explained variation in program condition over and above that accounted for by test-taking behaviors. The approximate degree of accuracy with which the two classes of variables predicted to program condition was examined by discriminant function analysis. When verbal and performance IQ scores were used for data, children were classified into treatment and control-condition groups with approximately 65% accuracy. When test-taking behaviors were used as data, children were classified into these two groups with approximately 78% accuracy. The findings of this study indicated that given the testing situation treatment-condition children exhibited significantly more work and verbal responses and less substitutive responses than the control-condition children; differences in IQ between these two groups were also significant. However, test-taking behaviors differentiated between groups more accurately. Based on these results, it would appear that educational intervention programs which rely exclusively on standardized tests as outcome measures may be ignoring a valuable aspect of program impact if children's behavioral repertoires are not also included in the evaluation.

Description

Keywords

Citation