versión On-line ISSN 1668-7027
MERINO SOTO, César y MUNOZ VALERA, Paola. A preliminary study on the socioeconomic impact on the scores of a set of multidimensional aptitude tests in pre-school children. Interdisciplinaria [online]. 2007, vol.24, n.2, pp. 161-184. ISSN 1668-7027.
The present study explores the relationship between socioeconomic level, gender and type of pre-school education center, and cognitive development in 323 pre-schoolers in Lima (Perú), using a multidimensional aptitude test. By applying the multivariate technique, we found statistical significant differences and medium size effects on the cognitive scores due to the educational center location, whereas gender and type of educational center differences were not found. The results were considered taking into account reliability levels and also according to the definition of socioeconomic levels. The differences found are discussed based on other articles as well as on the usefulness of the multidimensional aptitude tests for pre-schoolers. The sample consisted of 323 children of 5 years old, 167 boys and 156 girls, who go to both private and public pre-schools in a district of Lima. The proportion of public pre-school students (67.2%) was double the amount of private pre-school students (32.8%). Moreover, the private pre-schools had been recently formed (4 to 5 years before) and had children from families with a medium or medium to low socioeconomic level. In order to assess general and relevant aspects of pre-school learning, we applied a pre-school diagnostic test (De la Cruz, 1991), which can be used at a pre-school level (4-5 year-olds). This test contains the following areas: Verbal concepts, Quantitative concepts and Auditive memory, Visual perceptual aptitude and Visual-motor coordination. It is a potentia test because the time varies from one student to another so as to allow every child to answer every item. Furthermore, it is also a maximum performance test because the student tries to obtain the best possible score. We contacted several educational centers (private and public) from three different districts in Lima which accepted to participate in this study. The time of application of the test varied considering: individual skills, the number of children in the collective application, the presence of a teacher's aide for the collective application, and the attitude and motivation of the students. We used MANOVA to analyze profiles, a post hoc univariate contrast, and effect size estimations such as Cohen's d. Our study was based on three hypotheses: The first hypothesis was on the equality of levels (variability between subjects) that explores the differences that exist between both groups separately. As a result, we found no correlation between test performance and gender or school location. However, when considering type of school (public or private), we did find significant differences in one scale: Auditive memory, which was higher in public schools. The second hypothesis was the equality of dispersion (within-group differences), to identify if the answer variables have the same answers on average. We found similarities regarding school conditions (public or private); but there were small differences between gender and school location. The third hypothesis was regarding parallelism, which tests whether the pattern of elevations on the profile is similar between the two groups or not. We did not find important statistical dispersions and the effect size was small. The results coincide with other studies on the correlation between gender and socioeconomic level, and cognitive development. Among some of the limitations of our study, we list the following: for example, we used the school location to determine the children's socioeconomic level. Therefore, there were also extreme scores (outliers), which could vary the central tendency estimations and the dispersion. However, they were not trimmed, since in a previous unreported analysis without these data, researchers found only minimum changes. We also had occasional difficulties in applying the tests, due to the children's disposition to participate as well as their reaction concerning the test situation, which could have added random error on the results. However, we presume their negative reactivity was reduced because of the playful manner in which the tests were presented. The results of our research shed normative information regarding the performance of preschoolers due to gender and location of their educational centers. It is important to emphasize correlations as well as effect size of the results so as not to reach erroneous conclusions.
Palabras llave : Profile analysis; Pre-schooler; Aptitudes; Socioeconomic level; Effect size.