Explaining rating scale usage: The Semantic Threshold of Induced Categories.

摘要

The Semantic Threshold of Induced Categories model 1 (STIC1) is introduced as the cognitive process by which people assign observations to rating scale categories. STIC1 contends that rating scales prevent the respondent from using the full complement of descriptors available in the spoken language. When the number of categories are limited in this way, the descriptors that would have been available in the spoken language are annexed into the nearest adjacent rating scale category. This study describes a method to scale descriptors of frequency onto a continuum of frequency as it is used in the spoken language. The continuum was demonstrated to be stable across several relevant groups (gender and ethnicity) and conditions (within item and across item presentation order). Thresholds predicted by the STIC1 model adequately matched most of the 38 observed thresholds from 11 surveys with frequency-based ratings scales.