Abstract
The effects of quasi-polar and concept-scale interaction on the reliability of the semantic differential were investigated. Using a true experimental design one group of 150 out of a total of 300 randomly selected college students rated 10 pretested and standardized interactive and noninteractive concepts over 10 tested and verified bipolar scales. The second group of 150 subjects rated the same 10 concepts over 10 tested and verified quasi-polar scales. Results indicate that neither single rater reliability (range .462 to .532) nor group reliability ratings (range .972 to .982) for the semantic differential are affected ( p = .05) by quasi-polar scales or concept-scale interaction. While the semantic differential is shown to be extraordinarily robust in a variety of situations likely to be incurred in practice, the importance of reliability for the semantic differential, in light of the results, is questioned.

This publication has 3 references indexed in Scilit: