ABSTRACT
Periodically, the uses and misuses of probability statistics in social and behavioral science research have been reviewed. For instance, Don Lewis and C. J. Burke pointed to several misuses of the Chi-square test and an article by Hanan C. Selvin stimulated discussion on the general question of using statistics in social surveys. James K. Skipper, Anthony L. Guenther, and Gilbert Nass reviewed the discussion of substantive interpretation associated with significant levels. If the phenomenon of possible nonlinearity is not taken into account or if the implication of linearity is made in interpreting Chi-square, serious consequences again arise in interpreting data and in developing explanatory theories. The data were derived from major sociological journals published between 1955 and 1965 in a systematic search for three by three tables, both variables ordinal.
