ALEMU, BESUFEKAD
CARRANZA, LUIS
PEREZ, CHRISTIAN
Understanding Differences in Children's Test Scores Across Socioeconomic Status and Race
Working Paper, AEA Summer Training Program 2009, Massachusetts Institute of Technology, August 2009
Cohort(s): Children of the NLSY79
ID Number: 6626
Publisher: American Economic Association

Permission to reprint the abstract has not been received from the publisher.

In order to focus on how SES affects the gap in test scores over time, we must give a brief overview of the potential reasons for the gap in test scores in general. It could be possible that children with genetically superior ability perform relatively better on ability tests than those with less ability. Psychologists before the 1970s argued that genetic differences are not primarily responsible for test score performance since age scores of children vary with their respective ages (Kleinberg 1963). Many arguments have been made that it is school quality that affects children's test scores (Loehlin, Lindzey, and Spuhler 1975). Phillips et. al (1998), focusing on black children, argues that it is in fact home environment that contributes to lower tests scores, causing those with poor home environments to start school at a disadvantage with regards to tests scores. Fryer and Levitt provide evidence against this by finding that there is no gap at entry between black and white children, implying that it is in fact something about the school environment that induces such gaps. The consensus within the economic and psychologist literature appears to be aligned more with the view of Phillips than of Fryer and Levitt. We aim to provide evidence for or against Fryer and Levitt's findings by trying to replicate their results using a different data set. Our initial methodology will bear high resemblance to their original paper. We use pooled cross-sections to construct trends in tests scores among various age groups. But our work extends the research done by Fryer and Levitt by tracking the gap in scores at five different instances in time, as opposed to two instances. We also extend on their research by studying transitory income shocks as they relate to test scores via first differences models. Unfortunately, first differences models remove a permanent notion of SES precisely because SES is defined as time-invariant. We instead look at how SES and race affect cha nges in test scores and let the effect of income shocks differ by SES and race by including interaction terms. In summary, we use pooled cross-sections to analyze permanent income measures similar to the analysis in Fryer and Levitt and first difference models to analyze transitory income shocks as they relate to changes in test scores.

Search Start:  03:53:27
Search Finish: 03:53:27

URL: http://www.nlsinfo.org
Phone: (202) 691-7410
Fax: (202) 691-7425
NLS questions: usersvc@chrr.osu.edu