Supervisor: Olev Must, PhD, Associate professor
University of Tartu
Opponent: Jari-Erik Nurmi, PhD, Professor
University of Jyväskylä
The thesis examines how students' motivation and self-efficacy are related with their educational achievement in different countries. The thesis focuses on an attempt to answer the question that has risen in cross- cultural research recently - why the aggregated national means of educational achievement tests are strongly negatively correlated with the aggregated motivation to learn?
All of the studies in the thesis used PISA 2006 as a sample. The data from 55 countries were analyzed, the sample size was approximately 350,000. These data are a classic example of multilevel data and need appropriate methodological treatment. Measurements that use multi-group data with the aim to make conclusions for individual group members should consider an additional element in the observed score, namely a group-specific component. To estimate the nature of biases from group belongings, the estimation of measurement invariance is needed. I showed that PISA basic measures are invariant across nations in the configural and metric sense (same structure and functional relationships). But the measurements did not have equal intercepts - students equal in latent trait do not get the same observed test scores. Or in other words, the difference between the observed score and the true score is not only a random error, but something else. And this "something else" could play an important role in the calculations of group means.
Of course, belonging to some group and the possibility that the observed score includes something from group-belongings is not a reason for the negative correlations between group means. For a negative correlation the values of one variable should increase and the values of the other variable should decrease. And really, if average tests scores rise and the expression of learning motivation decreases across nations, the correlation will be negative. I showed that the unexpected negative association disappears if you consider the developmental level of a society. Why? One explanation could be that students' motivation to learn is not static. When expressing their motivational beliefs, students may use for reference the broad surrounding context. Environment, rich and full of opportunities, could suppress students' motivation to learn.
The difference in educational achievement by gender was greater in the countries whose indicators of development rank higher than those of other countries in the survey. In other words, the more developed a country was, the bigger the differences in educational achievement test results by gender were.