Social Econ Blog

Kennedy School professor Brian Jacob and co-author Lars Lefgren try and answer this question. Using data on parents' requests for teachers, a metric of teacher's value added, and principal's teacher evaluations from a midsize school district in the West, they find:

Families with children in higher poverty and minority schools in the district strongly value student achievement. When they make requests, they are more likely to pick teachers who provide high "value-added" in terms of student achievement scores and teachers whom the principal rates highly in terms of factors such as organization, classroom management, and enhancing student achievement. However, these parents were essentially indifferent to the principal's report of a teacher's ability to promote student satisfaction. Interestingly, the results are exactly reversed for families in higher-income schools. These parents are most likely to request teachers whom the principal describes as "a good role model" and/or good at promoting student satisfaction. They do not choose teachers who provide high "value-added" in terms of student achievement, or who receive high scores in this area from their principal.

The authors suggest several potential explanations for this finding. First, they note that education should be viewed as a consumption good as well as an investment good, and that it is possible that wealthier parents simply place a higher premium on the consumption value of schooling. Second, the authors note that these findings are consistent with a declining marginal utility of achievement on the part of parents. In other words, wealthier parents may believe that their children already have something of a head start in basic reading and math skills, so they value a strictly achievement-oriented teacher less highly than more disadvantaged parents whose children may not have these basic skills. More generally, these results suggest that what parents want from school is likely to depend on family circumstances as well as on parent preferences.

It would be interesting to do something similar here. I would love to know the relationship between student background characteristics and the type of classes they choose and how they rate courses for the CUE guide.