The Contributions of Paul Holland

Abstract

Paul Holland’s work over his long and varied career has shown both breadth and depth. He has made major contributions to the analysis of discrete data, to the study of social networks, to equating, to differential item functioning (DIF), to item response theory (IRT), and to causal inference. He has worked on a wide variety of applied problems ranging from scanner accuracy to test security to summarization of data on candidates. Any review of his contributions will necessarily provide a rather limited indication of his achievements. Nonetheless, several instructive themes can be found in his work. One is the long-standing connection with the analysis of discrete data. A second is a longstanding connection to the social and behavioral sciences. A third is an emphasis on the observed over the unobserved in the analysis of data. These themes interact and have been demonstrated in Paul’s work at least since graduate school. Paul’s doctoral dissertation concerned a new minimum chi-square test. His involvement in research in the social sciences reflects both his family background and his early association with his dissertation advisor Patrick Suppes (Robinson, 2005). The emphasis on the observed can be seen in his emphasis on observed-score equating and log-linear models rather than on latent-structure models, although Paul has made major contributions to IRT.