Ten Statisticians and Their Impacts for Psychologists

Abstract

ABSTRACT-Although psychologists frequently use statistical procedures, they are often unaware of the statisticians most associated with these procedures. Learning more about the people will aid understanding of the techniques. In this article, I present a list of 10 prominent statisticians: David Cox, Bradley Efron, Ronald Fisher, Leo Goodman, John Nelder, Jerzy Neyman, Karl Pearson, Donald Rubin, Robert Tibshirani, and John Tukey. I then discuss their key contributions and impact for psychology, as well as some aspects of their nonacademic lives. 3The disciplines of psychology and statistics have grown up as close companions; they are "inextricably bound together" (Stigler, 1992, p. 60). Although each can track their intellectual genealogies back millennia, as academic disciplines they are only about 100 years old. Early on, single individuals like Galton and Spearman could make substantive impacts in the core of both fields, but academic specialization and the development of both disciplines make this more difficult in modern times. Many psychologists like Abelson, Cohen, Cronbach, and Meehl have crossed into statistics, but their efforts have been mainly to bring statistical science into psychology. Others like Cattell, Coombs, Luce, Shepard, Stevens, Swets, Thurstone, and Tversky have focused on measurement and psychometrics. Measurement is fundamental for science (Borsboom, 2006), and psychologists have arguably done more for measurement theory than any other discipline (Hand, 2004). The people listed above would be on a list of psychologists worth knowing.The focus of this article is to compile a list of 10 statisticians who have done important work that psychologists should be aware of. The fact that psychologists take statistics courses throughout undergraduate and graduate training and that every empirical article includes statistics are testament to the importance of statistics within psychology.Unfortunately, statistical techniques are often taught as if they were brought down from some statistical mount only to magically appear in SPSS. This gives a false impression of statistics. Statistics is a rapidly evolving discipline with colorful characters still arguing about some of the core aspects of their discipline.The aim of this article is to present a list of the 10 statisticians whom I think psychologists should know about. By statistician, I am referring to people who would 4 likely be in a statistics department in the 21st century. I am not listing names prior to 1900, though interested readers should consult Stigler (1986). Limiting the list to 10 was difficult, and I based the list in part on personal preferences. The list is a mix of founding fathers and more recent stars, but the statistical advances had to be ones of importance to psychologists. The list is all male-however, the gender imbalance has been found in It is worth noting that the degrees of freedom he advised for χ 2 was in error (nk-1 for an n × k contingency table and the null of no association) and was later corrected by Fisher....

Bayesian Versus Orthodox Statistics: Which Side Are You On?

“…The question has the virtue of simplicity. Fisher argued that the only possible answers were ''yes'' (after a significant result, one can conclude there is a difference) or ''withhold judgment'' (after a nonsignificant result; see Baguley, in press;Dienes, 2008;Wright, 2010, for overviews of Fisherian and Neyman Pearson inference). A nonsignificant result does not allow a definitive conclusion, because there might be a population difference that the test was not sensitive enough to pick up.…”

Statistical techniques for juror and jury research

“…The group*belief means the model includes the interaction between the variables and their main effects. The difference between this model and the model with only the variables' main effects is: χ 2 (1) = 3.61, p = .06, a significance level Efron and Tibshirani (1993, p. 204; see also Wright, 2009) call ‘borderline evidence’ against the null hypothesis. The results are best illustrated graphically.…”

A Regression Framework for Effect Size Assessments in Longitudinal Modeling of Group Differences

“…Generalized linear models (Nelder & Wedderburn, 1972) is an extension of GLM that integrates OLS regression with logistic and Poisson regression analyses and uses maximum-likelihood estimation in analysis of both normally-distributed and categorical dependent variables, with similar structural models (Agresti, 2002; Hosmer & Lemeshow, 2000). Generalized linear models also subsumes GMA with fixed effects, including latent class growth analysis, and its formulation has recently been hailed as one of statistics’ most important contributions to psychology (Wright, 2009). …”