Chameleon effects in homework research: The homework-achievement association depends on the measures used and the level of analysis chosen

Using a data set specifically tailored to homework research, with a sample of 1275 students from 70 classes in Switzerland, the association between homework and achievement in French as a second language was tested at three levels (class level, between-student level, and within-student level). The strength and direction of the homework–achievement association depended on the homework indicator chosen and differed to some degree across analytical levels. At the class level, achievement was higher in classes set frequent homework assignments and in classes where students reported low overall levels of negative emotions when doing homework. At the between-student level, high individual homework effort and low levels of negative homework emotions predicted favorable developments in French achievement, whereas high homework time predicted lower achievement. At the intraindividual level, high homework effort, high homework time, and low levels of negative homework emotions were statistically significantly associated with positive student evaluations of the learning gains from the specific assignment.