Frequency effects and amplitude of category scale

Judgment of stimulus intensities with skewed frequences are affected by variations of context. If stimulus intensities are rated by using a category scale, then contextual biases are reduced by increasing the number of categories. Even if there are models which explain contextual biases due to the method to present stimuli, as Parducci's and Haubensak's model, it is not yet clear the nature of the interaction between the method to present stimuli and the variation of scale amplitude. In this experiment, squares with different brightnesses were presented with negatively or positively skewed ditributions of frequency. One group of participants rated the brightnesses using a scale with five categories (from 1 to 5), while another group could select all numbers included between 1 and 5. Results show that scale enlargement reduces contextual biases in subjective judgments. A t-test and a regression analysis show that the reduction of contextual biases is due to the "caution" principle, according to which participants prefer selecting central categories for the first stimulus and then categories which are near the central categories when a new stimulus appears. Therefore, when the scale has few categories, participants are forced to use up the extreme categories for the initial presentations increasing, therefore, the contextual biases in judgments.