No abstract

21 Ss performed complete ratio estimation of 7 stimuli in each of 6 continua. Individual scales were constructed and condensed into a measure of scale range. The coefficients of correlation between continua were positive and mostly statistically significant but much lower than the coefficients of reliability. The results support a two-factor hypothesis, according to which individual differences in scaling behavior reflect both a genuine perceptual variability and differences in response bias.

Data for 8 Ss were obtained with a modified technique for a category scale, a ratio estimation scale, and a free number production scale. When these are compared with previous data of Ekman and Rosner, the need for further research is clear.