< Back to previous page
Project
Interest and option choice in a developmental perspective. A multilevel multinomial modelling approach.
Maarten Pinxten, The Reciprocal Effects Model and educational choice. Alongitudinal study on academic self-concept, achievement and (post)secondary educational choice.
Doctoral dissertation submitted to obtain the degree of Doctor in Educational Sciences, 2013.
Supervisor: Prof. Dr. Bieke De Fraine Co-supervisors: Prof. Dr. Wim Van Den Noortgate
Prof. Dr. Jan Van Damme
This doctoral dissertation consists of two sections. In the first part, Manuscripts 13, we focus on the role of academic self-concept in explaining student achievement. The second part, Manuscripts 4 and 5, centers around educational choice.
The academic self-concept is generally considered both as an important predictor of academic achievement and as a valuable outcome variable in itself. The central focus of the first part of this doctoraldissertation is on the longitudinal relation between academic self-concept and achievement and four aspects of this relation were explored in greater detail. The results of Manuscript 1 showed that there exists a certain developmental trend</> in the longitudinal relation between both variables in a way that only by the end of secondary education, this relation becomes reciprocal. Thus, supporting the premises of the Reciprocal Effects Model (REM), academic self-concept positively affected achievement and vice versa. Furthermore, small gender differences</> were observed in this longitudinal relation. However, given the small effect sizes, the results of this first study should be interpreted carefully in the light of the articulated methodological limitations. Manuscripts 2 and 3 center around the delicate balance between conceptualization of psychological constructs on the one hand and their measurement on the other. The results of Manuscript 2 clearly showed that the type of measurement of academic achievement</> (standardized tests versus teacher ratings) substantially influenced the causal pattern of cross paths observed in REM studies. Thus, the type of achievement indicatorused moderates the relation between academic self-concept and achievement. Different conceptual connotations of tests and teacher ratings affect the outcomes of the quantitative modeling of REM relationships. A similar conceptual issue emerged in Manuscript 3. The multidimensionality ofacademic self-concept in terms of domain-specificity has been widely accepted but its multidimensionality in terms of motivational properties (the so-called affect-competence distinction</>) is still under debate. In evaluating ones own subject-specific or general academic abilities, both affective reactions (e.g., enjoyment) and cognitive evaluations (i.e., rational competence evaluation) are at stake. The question then arises whether these affective responses should be rather considered as a subcomponent of the academic self-concept or as a separate latent construct. Based on both a within-network approach (confirmatory factor analysis) and a between-network approach (relations with external outcomevariables), the results of Manuscript 3 clearly argue in favor of the latter. Math enjoyment and math competence could hardly be considered as combined indicators of a higher order factor math self-concept. Additionally, differential longitudinal relations between both constructs on the one hand and math achievement and math effort expenditure were observed.
In the second part of this dissertation, we tried to gain insight into the socio-cognitive determinants of the choice of study program in the academic track of secondary education (Manuscript 4) and the choice of a major in the first year of university (Manuscript 5). Different theoretical models stipulate educational choice processes as being complex and multifaceted phenomena wherein a large number of student and school level factors come into play (e.g., gender, SES, prior achievement, prior subject uptake, occupational interests, future aspirations and self-concepts of ones abilities). On the basis of multilevel and ordinary multinomial logistic regression (in Manuscripts 4 and 5 respectively), the joint effect of each of the aforementioned variables on the type of study program chosen was thoroughly investigated. The results of our analyses clearly showed that almost all of these variables significantly contributed to the choice of a particular study program. Choice of astudy program in secondary education and major selection in university still followed a gender-stereotypical choice pattern with girls and boysrather opting for language/humanities and science/mathematics majors respectively. Different interest patterns (especially the higher interest in science of boys) between boys and girls accounted for a substantial part of this gender-stereotypical pattern. Furthermore, negative choice mechanisms were identified in secondary education in a way that students with a lower SES were less inclined to choose a study program with Latinor Greek, even after controlling for prior achievement and interests. Similarly, it seemed that students opting for a humanities study program were primarily pushed away from the other study programs based on lower achievement rather than an increased interest for humanities. Finally, students SES did not influence the type of major chose in the first yearof university.
Doctoral dissertation submitted to obtain the degree of Doctor in Educational Sciences, 2013.
Supervisor: Prof. Dr. Bieke De Fraine Co-supervisors: Prof. Dr. Wim Van Den Noortgate
Prof. Dr. Jan Van Damme
This doctoral dissertation consists of two sections. In the first part, Manuscripts 13, we focus on the role of academic self-concept in explaining student achievement. The second part, Manuscripts 4 and 5, centers around educational choice.
The academic self-concept is generally considered both as an important predictor of academic achievement and as a valuable outcome variable in itself. The central focus of the first part of this doctoraldissertation is on the longitudinal relation between academic self-concept and achievement and four aspects of this relation were explored in greater detail. The results of Manuscript 1 showed that there exists a certain developmental trend</> in the longitudinal relation between both variables in a way that only by the end of secondary education, this relation becomes reciprocal. Thus, supporting the premises of the Reciprocal Effects Model (REM), academic self-concept positively affected achievement and vice versa. Furthermore, small gender differences</> were observed in this longitudinal relation. However, given the small effect sizes, the results of this first study should be interpreted carefully in the light of the articulated methodological limitations. Manuscripts 2 and 3 center around the delicate balance between conceptualization of psychological constructs on the one hand and their measurement on the other. The results of Manuscript 2 clearly showed that the type of measurement of academic achievement</> (standardized tests versus teacher ratings) substantially influenced the causal pattern of cross paths observed in REM studies. Thus, the type of achievement indicatorused moderates the relation between academic self-concept and achievement. Different conceptual connotations of tests and teacher ratings affect the outcomes of the quantitative modeling of REM relationships. A similar conceptual issue emerged in Manuscript 3. The multidimensionality ofacademic self-concept in terms of domain-specificity has been widely accepted but its multidimensionality in terms of motivational properties (the so-called affect-competence distinction</>) is still under debate. In evaluating ones own subject-specific or general academic abilities, both affective reactions (e.g., enjoyment) and cognitive evaluations (i.e., rational competence evaluation) are at stake. The question then arises whether these affective responses should be rather considered as a subcomponent of the academic self-concept or as a separate latent construct. Based on both a within-network approach (confirmatory factor analysis) and a between-network approach (relations with external outcomevariables), the results of Manuscript 3 clearly argue in favor of the latter. Math enjoyment and math competence could hardly be considered as combined indicators of a higher order factor math self-concept. Additionally, differential longitudinal relations between both constructs on the one hand and math achievement and math effort expenditure were observed.
In the second part of this dissertation, we tried to gain insight into the socio-cognitive determinants of the choice of study program in the academic track of secondary education (Manuscript 4) and the choice of a major in the first year of university (Manuscript 5). Different theoretical models stipulate educational choice processes as being complex and multifaceted phenomena wherein a large number of student and school level factors come into play (e.g., gender, SES, prior achievement, prior subject uptake, occupational interests, future aspirations and self-concepts of ones abilities). On the basis of multilevel and ordinary multinomial logistic regression (in Manuscripts 4 and 5 respectively), the joint effect of each of the aforementioned variables on the type of study program chosen was thoroughly investigated. The results of our analyses clearly showed that almost all of these variables significantly contributed to the choice of a particular study program. Choice of astudy program in secondary education and major selection in university still followed a gender-stereotypical choice pattern with girls and boysrather opting for language/humanities and science/mathematics majors respectively. Different interest patterns (especially the higher interest in science of boys) between boys and girls accounted for a substantial part of this gender-stereotypical pattern. Furthermore, negative choice mechanisms were identified in secondary education in a way that students with a lower SES were less inclined to choose a study program with Latinor Greek, even after controlling for prior achievement and interests. Similarly, it seemed that students opting for a humanities study program were primarily pushed away from the other study programs based on lower achievement rather than an increased interest for humanities. Finally, students SES did not influence the type of major chose in the first yearof university.
Date:1 Oct 2009 → 24 May 2013
Keywords:Academic self-concept, Structural Equation Modeling (SEM), Multinomial, Multilevel, Choice
Disciplines:Education curriculum, Education systems, General pedagogical and educational sciences, Specialist studies in education, Other pedagogical and educational sciences, Psychological methods, Mathematical and quantitative methods, Social theory and sociological methods, Political theory and methodology, Parenting and family education
Project type:PhD project