Randomized Scales, Scale-Checking Styles, & Reaction Time: Methodological Implications.

This empirical study examines whether scale-checking habits cause differences among responses to three types of semantic differential scales: a scale in which all the favorable adjectives were on the left side of the scale; a scale in which all the favorable adjectives were on the right side of the scale; and a scale in which the right-to-left and left-to-right scales were randomly mixed. The questionnaire consisted of fifteen items designed to rate customer attitudes toward three local malls. One additional issue of interest was to test if the reaction time to each of these scales was significantly different. As expected, the reaction time for the random scale displayed more variability around the mean. The type of scale used had a definite and statistically significant effect on the responses. The L-R scale biased the responses towards the favorable adjective, the R-L scale generated a bias in the opposite direction, and the mixed scale successfully eliminated any such bias. The implications of the findings are of interest to marketing scholars and practitioners that regularly use semantic differential scales and are concerned about the psychometric properties and consequent validity of the results obtained. [ABSTRACT FROM AUTHOR]