A Bibliography of Multiple Comparison Procedure

A comprehensive bibliography of multiple comparison procedures includes 60 entries spanning from 1935 to 1993. The earliest entry is R.A. Fisher's 1935 book, "The Design of Experiments," published by Oliver and Boyd, which laid foundational concepts in experimental design. Following this, J.W. Tukey's 1949 article, "Comparing individual means in the analysis of variance," published in Biometrics, marked a significant advancement in statistical methods. D.B. Duncan contributed several important works, including his 1951 article on significance tests for ranked treatments and a 1955 piece on multiple range and F tests, both of which have been referenced multiple times in later studies.

In 1953, Scheffe proposed a method for judging all contrasts in analysis of variance, which was published in Biometrika, alongside Tukey's influential work, "The Problem of Multiple Comparisons," noted for its widespread use in statistical literature. The 1960s saw a surge in contributions, with significant entries like Dunn's 1961 article on multiple comparisons among means, which has been referenced extensively. The decade also included notable works by Steel, who published multiple articles on rank sum tests and comparisons, further enriching the field.

The trend of increasing scholarly output continued into the 1970s and 1980s, with 15 entries in 1975 alone, including works that explored various aspects of multiple comparisons in psychological research and statistical methodology. By 1982, the bibliography recorded 12 entries, reflecting ongoing developments in the field, including contributions that incorporated Bayesian approaches, such as Duncan's 1965 article on Bayesian methods.

The chart section reveals a clear trend in the inclusion of specific terms in titles over the years. For instance, in 1965, there were six total entries, with one explicitly mentioning "Bayes" in the title. The following year, 1966, maintained a similar count, with three entries. The late 1960s and early 1970s witnessed a notable increase in the number of works referencing "False Discovery Rate" or "FDR," culminating in a peak of 17 entries in 1976, though the use of Bayesian terminology remained relatively rare until the 1980s.

By 1993, the bibliography included 16 entries, with one work explicitly referencing "FDR." This reflects a broader trend in statistical literature towards addressing the complexities of multiple comparisons, including the challenges of controlling error rates in hypothesis testing. The evolution of these methodologies highlights the dynamic nature of statistical research and the ongoing dialogue among scholars in the field.