58 Factors associated with differences in digitally measured student perceptions of teaching quality or are classes with high mathematics grades less demanding regarding teaching quality, because the students in those classes are bright? Based on our research, the aforementioned questions cannot be answered unequivocally, because our research does not allow for causality claims. That would require more objective evaluations of teaching quality, for example, by trained lesson observers observing teachers’ lessons. Other research on causality claims might help in answering these questions. A study by van de Grift et al. (2011) showed that more experienced teachers (up to 15 years of teaching experience) receive higher teaching quality ratings from external observers. Furthermore, many studies have shown the positive correlation of student achievement and effective teaching practices (e.g., Creemers, 1994; Day et al., 2008; Hattie, 2008; Marzano, 2003; Muijs et al., 2014; Pianta & Hamre, 2009; Reynolds et al., 2014; Rosenshine, 1995, 2012; Sammons et al., 1995). 3.5.2 Unrelated factors In this study, seven of the variables investigated did not seem to be related to bias. However, in previous research (see the theoretical framework), these factors were related to differences in students’ ratings of teaching quality. A possible reason for the differences in findings between this study and earlier studies may be that some of the variables in our study had very little variance in our sample. For example, the classes studied all consisted of about 25 students (which is normal in the Netherlands), and the number of ethnicities in those classes was small. In the teacher sample, there was also limited variability in gender and age. This might be a reason why these variables did not show up with an effect, while, for example, teacher likeability did. Furthermore, in other educational contexts such as higher education, students are normally older than the students in the classes of the current study. The interplay between teacher and student gender might be more of an issue for older students, and could cause bias. The distance between students and their lecturers is probably higher in higher education than in secondary schools, which might also have caused differences between the outcomes of the studies. Another possible reason why some variables did not seem to cause bias in the current study could be that the items in the Impact! tool were formulated in a lessonspecific way (what happened in a specific lesson) and were not about the general quality of the teacher’s instruction (which was the case in the student perception questionnaires used in other studies). Maybe the differences in item formulations contributed to the differences in the findings between the studies.
RkJQdWJsaXNoZXIy MjY0ODMw