Numerous conceptual and statistical difficulties complicate the use of raw thought frequencies in cognitive assessment. To overcome such problems, investigators have used a variety of ratio scores; however, these are not easily comparable across studies and their use evokes a variety of associated difficulties. Therefore, the present investigation compared results using raw frequencies and a variety of ratio scores on a cognitive assessment task where both the valence and the focus of attention of thoughts were examined. Scores evaluated included a positive/negative ratio as well as Schwartz's States of Mind (SOM) ratio [positive/(positive +negative)] with and without correction of zero frequencies of either positive or negative thoughts. Results indicate that (a) partitioning thoughts as a function of attentional focus (self, other, situation-focused thoughts) yield different frequencies and ratios for each thought type, (b) positive and negative thoughts are independent, (c) although different ratios and correlations are appropriate, depending on the nature of the assessment task, results using the SOM ratio with a correction factor most closely resemble results obtained when using frequencies, (d) when a positive/negative ratio is used, a correction factor and a log transformation reduce the effects of missing data and of positive skewness, and (e) the mean of subjects' SOM ratios and the ratio of subjects' mean frequencies yield similar results. © 1990 Plenum Publishing Corporation.
