PerceptualandMotor Skills, 1990, 71, 335-338. O Perceptual and Motor Skills 1990

SYLLOGISTIC REASONING IN PUERTO RICAN BILINGUAL ELEMENTARY SCHOOL CHILDREN ',' MATTHIJS KOOPMANS Yesbiva University Summary.-The purpose of the present study was to measure the effects of language proficiency and grade (age) on the reasoning of Puerto Rican bilingual children in elementary school. 20 syllogisms were administered, 10 in English and 10 in Spanish, and accuracy of and strategy for solution were examined. Solutions were more often accurate, and the proportion of theoretical explanations given for them was higher in Spanish. Differences by grade were not significant except a higher proportion of theoretical explanations were given by children in Grade 5 for syllogisms in Spanish.

The purpose of the present study was to investigate the extent to which the syllogistic reasoning of Hispanic bilingual elementary school children is affected by their proficiency in the language in which the reasoning task is administered. Previous studies have established that for a bilingual adult population, there is little difference in the ability to solve reasoning tasks in the weaker language, although the speed with which the solutions are generated is slower in the second language (dlAnglejean, Gagnon, Tucker, & Winsberg, 1979; Duran, 1981, 1988). Moreover, it appears that solutions produced by bhngual adults are less accurate when the task is administered in the second language, although subjects are equally likely to understand the premises in the two languages (d'Anglejean, et al., 1979). This latter finding is consistent with the results for monolingual adults which indicate that variations in reasoning performance are attributable to a great extent to the complexity of the premise information rather than reasoning ability as such (Guyote & Sternberg, 1981; Johnson-Laird & Steedman, 1978). Due to the formal nature of the operations to be performed on the premise information (the problem-solver needs to assume that the premise information is hypothetical to solve the syllogism successfully), some variations in reasoning abhty can be expected in elementary school children because children are in a transition from the concrete operational to the formal operational stage (Piaget, 1928). Orasanu and Scribner found that, for monolingual children, this transition indeed is reflected in their sylIogistic 'Part of this study has been funded by the Center of Language Education and Research under the auspices of the Office for Educational Improvement in a grant to Catherine Snow, Harvard Graduate School of Education. For their help in the various sta es of this project, I thank Kenji H a b t a , Kay HiU, Catherine Snow, Terry Tivnan, and John ~ d t t , Address correspondence to M. Koopmans, Ferkauf Graduate School of Psychology, Mazer Hall, Albert Einstein College of Medicine Campus, 1300 Morris Park Avenue, Bronx, NY 10461.

336

M. KOOPMANS

performance. Comparing second and fifth graders, both solution accuracy and proportion of theoretical justifications were higher in Grade 5 than in Grade 2. The differences by grade for both accuracy and justification type indicate that the effects of language proficiency on reasoning ability, found among bilingual adults, needs to be established separately for bilingual children. We also need to know whether differences between first and second languages are reflected both in solution accuracy and type of justification given. Whether the results obtained for monolingual elementary school children also apply for bhngual children, particularly in the second language, musr be assessed as well.

Subjects Subjects were 93 Puerto Rican elementary school children who were enrolled in the bilingual program at the public schools in New Haven, Connecticut. Children were individually taken out of the classroom for testing in English as well as in Spanish. One session was conducted exclusively in Spanish, the other exclusively in English. To balance effects of learning, subjects were randomly assigned to the 'English-first' or 'Spanish-first' group. Children from Grades 3 to 5 were included in the study, and it was established in advance on the basis of teachers' ratings, that subjects would be testable in both languages. To assess reasoning, 10 syllogisms were used for each testing session. To account for effects of ordering and level of difficulty, the syllogisms were administered in randomized order by shuffling the syllogism cards before each administration. Syllogisms were translated into Spanish by a team of Puerto Rican translators, and agreement was ensured for the translations. It was also established in advance that no uncommon words or idiom was used in either language. Organization of content terms and quantifiers, by which structure and complexity of syllogisms are determined (see, e .g., Johnson-Laird, 1983), was exactly the same in English and Spanish. ~

-

Criteria for Scoring Syllogistic reasoning was coded according to the following criteria, adapted from Orasanu and Scribner (1982): (1) solution accuracy: average proportion of correct answers, (2) solution strategy: average proportion of theoretical justifications. Interscorer reliability, computed for reasoning in English on the basis of 10 protocols, ranged from 68% to 89%. RESULTS Table 1 shows the scores in both English and Spanish for solution accu-

337

BILINGUAL CHILDREN'S SYLLOGISTIC REASONING

racy as well as justification type. Solution accuracy was higher for children in Grade 5 than for those in Grades 3 and 4 in both languages. O n e way analyses of variance indicate that differences between grades were not significant in either English or Spanish. Solution accuracy scores were also higher in Spanish than in English for all three grades. An analysis of variance with language of administration as a repeated measure, and grade as a betweensubjects variable indicates that the main effect of language of administration on solution accuracy is significant (F,,,, = 9.01, p< .01), whereas the main effect of grade level is not. The interaction of language by grade is also nonsignificant. TABLE 1 MEANSAND STANDARD DEVIATIONS BY GRADEFORACCURACY OF S O L ~ O N AND J U S T ~ C A T IOF O NSYLLOGISMS IN ENGLISH A N D SPANISH Language

Grade 3

Grade 4

Grade 5

P

English Solution Accuracy

M SD Justification

M SD Spanish Solution Accuracy

M SD Justification

M SD

Table 1 shows a similar trend for solution strategy. I n Grade 5 , subjects score higher than those in Grades 3 and 4, but the tendencies fall short of significance in English and in Spanish only reach the .10 level (F,,,, = 3.04, p < . l o ) . Multiple comparisons indicate that for solution strategy in Spanish, the difference between means for Grade 5 and the other two grades was significant (Tukey's LSD = 2.81, p = .05). A repeated-measures analysis of variance in which language was varied within subjects and differences among grades were also assessed, showed a significant effect for language of administration (F,,,, = 9.30, p

Syllogistic reasoning in Puerto Rican bilingual elementary school children.

The purpose of the present study was to measure the effects of language proficiency and grade (age) on the reasoning of Puerto Rican bilingual childre...
154KB Sizes 0 Downloads 0 Views