Statistical approach to the word association test

Six different quantitative single-valued measures for the response domain in free Word Association Test are summarized. The measures are w--the probability of the primary response; K--the number of different responses related to the number of subjects; H--the number of idiosyncratic responses related to the number of subjects; Z--the coefficient proposed by Zinger (1982); Ri--the coefficient of isonymy defined by Lasker (1977); and I--the entropy of the response distribution. These measures are compared on the theoretical level. The results are demonstrated on an experiment in which 175 subjects responded to a list of 40 different stimuli words. The list was presented twice. The measure I is preferred for general application from both theoretical and experimental points of view.