In this paper, we respond to J. C. Bradbury’s critique of our 2011 Labour Economics paper examining the peer effect of Jose Canseco. None of Bradbury’s criticisms have any merit, and many reveal a severe misunderstanding of basic econometrics. For example, Bradbury accuses us of not deleting enough years from the sample, not censoring the sample on an outcome measure, and not controlling for average performance measures for each year explicitly when we have already included dummy variables for each year. Bradbury claims that we distort our findings, but he overlooks the parts of our paper that do not fit his thesis. Bradbury reexamines the performance of Canseco’s teammates empirically and argues that our results are sensitive. However, this should not be surprising because Bradbury performs a completely different and highly flawed analysis. In particular, he fails to realize that he is estimating very different parameters which are difficult, if not impossible, to interpret. His specification and estimation are based on very restrictive assumptions which are not necessary, nor are they justified or even acknowledged. After examining every one of Bradbury’s attacks on our paper, we conclude that none provides a convincing reason to reject our conclusions.
This article is a response to Did Jose Canseco Really Improve the Performance of His Teammates by Spreading Steroids? A Critique of Gould and Kaplan by John Charles Bradbury (EJW, January 2013).