Load Home page + menu
Last update 26 march 2010

20th century


The 20th century is characterized by some major methodological battles. Firstly there is a disagreement with regard to the preference for correlational large scale studies (Karl Pearson) versus experimental small scale studies (Ronald Fisher). The field of experimental small scale studies witnesses the emergence of a second controversy: H0 testing (Ronald Fisher) versus the inclusion of H1 and the concept of Power (Jerzy Neyman & Egon Pearson).
Spearman's belief in one general intelligence (g) factor, which supposedly was the driving force behind the development of factor analysis, leads to arguments, which last for several decennia, with Thurstone and others who gradually look upon factor analysis as 'just' a way to simplify the data.
After the second world war non-parametric analysis booms and the invention of the computer creates numerous possibilities for implementing new (and old) ideas such as multidimensional scaling, bootstrapping and multivariate analysis.
1900 Karl Pearson presents the idea of the Chi square distribution
1904 Charles Spearman lays the basis for Factor analysis and completes it 8 years later. 
Spearman presents the correlation coefficient for rank data.
1908 William Gosset presents his work on the t distribution and the application of the t test.
The first appearances of the t test in psychology and related areas are in the early 30s, but Shen (1940) still refers to the t test as not yet in general use in the field of education
1925 Ronald Aylmer Fisher publishes his Statistical methods for research workers. It is the first textbook presentation of analysis of variance.
1933 Andrei Kolmogorov presents an axiomatic basis for probability in his book Grundbegriffe der Wahrscheinlichkeitsrechnung (Foundations of the Calculus of Probabilities). He also introduces the test statistic |Fn(x) - F0(x)|.
1933 Harold Hotelling publishes his work on principal component analysis
1939 Vladimir Smirnov uses the statistic developed by Kolmogorov to construct the Kolmogorov-Smirnov test
1945 Frank Wilcoxon presents the the rank-sum test and the signed-rank test
1976 Gene Glass publishes his article on combining the results of multiple studies and coins this approach meta-analysis. Although many of these ideas existed long before (Lush 1931, Fisher 1932; Pearson 1933, Snedecor 1946) it is Glass who gives this approach the impetus to achieve the status it enjoys since.
1977 John Tukey presents exploratory data analysis (EDA) as an antidote to the ritualized testing of hypotheses instead of first looking at the data