Testing the Reciprocal Effects Model in Migrant and Nonmigrant Student Samples

• In Event: Self-Efficacy, Self-Concept, and Their Relations to Academic Achievement in Minority and Migrant Groups: An International Perspective

Abstract

Objectives
We study reciprocal effects of domain-specific academic self-concepts and academic achievement in different subgroups of German students. While previous research has shown that reciprocal effects are invariant across different age groups and gender, there is not much evidence whether this is also true for migrant and non-migrant students.

Theoretical Framework
Calsyn and Kenny (1977) contrasted the self-enhancement and skill-development hypothesis. In the self-enhancement hypothesis, self-concept is considered to be a prerequisite of achievement. This may be caused by more engagement and effort in domain-specific activities, which then may result in higher achievement (Marsh & Yeung, 1997). In contrast, the skill-development hypothesis suggests that academic self-concept is a consequence of achievement, which can be explained by social comparisons (Möller, Pohlmann, Köller, & Marsh, 2009). To date, many researchers act on the assumption of an integrated model: the reciprocal effects model (Marsh & Craven, 2006). In this model, it is postulated that both are a cause and effect of each other. Much research (Marsh & Yeung, 1998) provided evidence that these reciprocal effects are invariant across gender as well as across different age groups. Not much work, however, has addressed the question whether the invariance assumption is also true across students with or without migration background. In Germany, Turkish students form a large group. Unfortunately, previous research has not analyzed the relations between self-concepts and achievement in these students.

Method
We collected repeated measurement data from two German samples of secondary students. The first data collection took place at the beginning of secondary school in grade 5, and the second data collection was at the beginning of grade 7. Sample 1 consisted of N = 1004 students (50.1 % female) and Sample 2 comprised N = 842 students (49.7 % female). Half of the Sample 1 (50.1%) had an immigration background, mostly Turkish. In Sample 2, 51.8% had an immigration background, again mostly Turkish.
Measures. Students’ math and reading self-concept was measured by means of 5 items that students had to answer on a four-point scale. Cronbach’s alphas were above .80. Reading and math achievement was assessed with standardized tests with sufficient curricular validity. Cronbach’s alphas ranged from .78 to .83.

Results
Multi-group structural equation modeling using Mplus 6.12 (Muthén & Muthén, 2010) was carried out to test the invariance of the reciprocal effects model across samples and groups of students (migrants vs. non-migrants). In both samples, the findings supported the reciprocal-effects model for both reading and math. Testing the model in migrant students and non-migrant students revealed no substantial differences suggesting that the model parameters were the same in both groups.

Significance
This study is one of the few to investigate the reciprocal effects model in a group of Turkish students. Results support that the model also holds in this group. Bearing in mind that most of the Turkish students in German schools show low achievement levels, the findings suggests that enhancing their achievement could be accompanied by measures that also support their self-concept.

Authors

• Olaf Koeller, Leibniz Institute for Science and Math Education

• Jan Retelsdorf, IPN - Leibniz Institute for Science and Mathematics Education