Response Set and the Measurement of Self-Designated Opinion Leadership

The internal consistency and discriminant validity of the items comprising the Rogers self-designating opinion leadership scale were investigated utilizing notions suggested by Campbell and Fiske's multitrait-multimethod matrix approach to establishing convergent and discriminant validity. While the items were shown to be highly intercorrelated, their ability to discriminate consistently between two types of opinion leaders was found to be limited and there were indications that trait-irrelevant method factors have a substantial effect on scores obtained on the scale. Acquiescence response set appeared to be present when the scale was used to identify opinion leaders for one type of activity, but did not appear operative when the scale was administered to the same sample with reference to a different category of behavior. This suggests that the types of response bias which may affect the Rogers scale and the extent of their influence depend on the area of opinion leadership to which the scale is applied. Randomizing the order in which response alternatives are presented and/or providing a neutral response category may help control some of the unwanted response bias.