Fishing in the Bay

Statistics features heavily in every psychology course. My son has just completed the rats and stats section of the course – and he vastly preferred the rats. Some psychology researchers, like Spearman, developed their own methods and have become household names in our field. But have statisticians influenced the field of psychology? We would certinaly hope so. The British Psychological Society agrees and have recently compiled an annotated list of the 10 statisticians (who were not psychologists) who have most influenced the field of psychology.**

Here is their list with a little commentary from me. If you want to have a look at the kind of statistics that psychology students learn see HERE.

· #1: Karl Pearson (1957-1936): regression to the mean, correlation coefficient, chi-square test of independence, classification of distributions (and racial eugenics….)
· #2: Ronald Fisher (1890-1962); ANOVA, exact tests, permutation tests. More generally responsible for most of statistical inference theory but perhaps slipped up on the causal effects of smoking.
· #3: Jerzy Neyman (1894-1981) foundations of sampling, Hypothesis testing, confidence intervals. Psychologists love hypothesis testing. It looks so damned scientific! I guess everyone knows that Fisher and Neyman/Person had contrasting views of statistical inference.
· #4: John Tukey (1915-2000) honest significant difference (range) test after ANOVA. Even though he is famous for a test, he more generally stressed exploratory statistical techniques over confirmatory ones (hypothesis testing).
· #5: Don Rubin (1943-) Meta-analysis and effect sizes. Amongst statisticians he is more famous for missing data methods and causal models.
· #6: Bradley Efron (1938-) for bootstrap. I am not actually aware that bootstrap is heavily used in psychology. Their experiments are usually well designed
· #7: Sir David Cox (1924-) for the Box-Cox transformation. What a tiny achievement to laud when considered against the edifice of his general contributions.
· #8: Leo Goodman (1928-) for measures of association in contingency tables. His gamma statistic is a commonly used measure of rank correlation, closely related to Kendall’s tau. I do not see how Kruskal did not get an equal mention here.
· #9: John Nelder (1924-) for the Generalised Linear Model. I do not really associate him with psychology at all; more with agricultural field trials at Rothampsteadand and the likelihood principle. I reckon the BPS are struggling to make up two hand’s worth at this stage.
· #10: Robert Tibshirani for LASSO. Do psychology researchers really use this?

Did anybody notice the absence of Bayesians? I am outraged! No really. Either this is a big oversight, or psychologists do not use these modern methods.

**Hat tip to Patty Solomon for alerting me to this.

You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.