Kumulative Dissertation, Otto-Friedrich-Universität Bamberg, 2021

The aim of this dissertation is to broaden our current knowledge of the effects of teachers and their qualifications, as key school inputs, on the human capital accumulation of kindergarten and primary school children in Ecuador, a developing economy, and Germany, a developed one. Even though these two countries are vastly different in their economic, scientific, social and cultural characteristics, their contexts are similar in that little research has been conducted there on teachers in the framework of the education production function, and particularly of research with causal interpretation.
This dissertation consists of an introduction and three independent scientific articles organized into Chapters.
Introductory Chapter 1 reviews theoretical and empirical literature on teachers as a school input affecting students’ human capital accumulation in the education production function framework.
In Chapter 2, I evaluate whether Ecuador's new teacher recruitment process, which requires teacher candidates to pass national entrance tests before they are allowed to participate in merit-based competitions for tenure at public schools, has served as an effective screening device and ultimately helped to improve student achievement in the first grades of primary school. To answer these questions, I analyze data from a unique Ecuadorian survey of schools in the academic year 2011-2012, which I match to individual teacher recruitment records from Ecuador’s Ministry of Education. I first estimate a value-added to student achievement model using Ordinary Least Squares (OLS). I then use Propensity Score Matching (PSM) to simulate a random assignment of students to teachers and estimate causal treatment effects. The evidence suggests that teachers awarded tenure through the new policy (test-screened tenured teachers) were no more effective overall in improving students' achievement in reading or math. Nonetheless, the OLS and PSM results also suggest that these teachers had a significant positive effect on reading achievement for students from poor households. The PSM estimations show an average treatment effect of about a 9 percent of a standard deviation in reading achievement for students living in poverty.
In Chapter 3, Guido Heineck, Yyannú Cruz-Aguayo and I implement an experimental evaluation of the impact of Ecuador's new teacher recruitment process for kindergarten. We link administrative teacher information from Ecuador’s Ministry of Education to data from a unique experimental study where almost 15,000 kindergarten children were randomly assigned to their teachers in the 2012-2013 school year. Our results show positive and significant effects of test-screened tenured teachers of at least an 11 percent of a standard deviation for language, and a 9 percent of a standard deviation for math. These effects persist even after controlling for teacher academic degree, experience, cognitive ability, personality traits and classroom practices. In addition, our estimations show that tenured teachers significantly outperform contract teachers in Ecuador, which contrasts with recent evidence on the positive effects of fixed-term contract teachers in other developing countries. Furthermore, we confirm that the effects of test-screened tenured teachers on language learning are stronger for vulnerable children who started the school year with lower baseline test scores or came from socioeconomically disadvantaged families.
Johanna Sophie Quis and I also study the effect of teachers on skill formation in Chapter 4, but in the context of a developed country, Germany. We use primary school data from the German National Educational Panel Study (NEPS) to estimate classroom effects on mathematical and language competence development, which are driven by teacher quality differences across classrooms. We estimate a value-added model with individual classroom fixed-, as well as random effects. Both model specifications apply empirical Bayes shrinkage to adjust the classroom effects’ estimates by their level of precision. Our results show substantial classroom effects and quality differences across the first grades of German primary school. One standard deviation increase in classroom effectiveness is associated with at least a 12 percent of a standard deviation increase in student mathematical competence scores, and at least 14 percent of a standard deviation increase in language competence scores. In addition, we find that none of the teacher characteristics typically used in teacher recruitment processes significantly explain the classroom quality differences. Interestingly, as parental assessment of teacher quality is the only indicator significantly associated with classroom effectiveness in language competence development, parents seem to be able to identify more effective language teachers.