Effects of Advanced Course-taking on Math and Science Achievement: Addressing Selection Bias Using Propensity Scores

Does taking advanced courses improve scores on basic achievement tests? This question is difficult to answer because there are many systematic differences between students who choose to take advanced courses and those who don’t. This paper uses cutting-edge propensity score methods to control for the selection bias that exists when studying the effects of advanced course-taking on a basic achievement test in math and science. In particular, the differential achievement between advanced course-taking and non-advanced course-taking students in a basic math and science proficiency test in high school is estimated, controlling for several background variables and using data from the Third International Mathematics and Science Study (TIMSS). Because propensity score methods adjust for selection bias on observed background variables, but not on unobserved background variables, we also consider the potential impact of selection bias due to unobserved measures (hidden bias) through a sensitivity analysis. Our analysis found that advanced course-taking is consistently associated with higher achievement. Sensitivity analysis suggests that this can be the result of hidden bias due to extreme differences in unobserved background variables, but cannot be explained if the differences in unobserved background variables are only moderate.