Comparison of Three Methods for Assessing Attitudes

Abstract

Conducting survey research involves making choices concerning the methods of gathering attitude data. Several methods are available, each with its advantages and disadvantages. Three methods were compared in this study. They included ratings, rankings, and paired comparisons. The responses of 392 undergraduate students were included in the analyses. Each participant completed a survey consisting of three attitude objects of five items each, one set formatted using a rating scale, another using rankings, and a third using paired comparisons. Three survey forms were developed so that each attitude object could be presented in each format. The three forms were presented randomly to approximately equal numbers of participants. Analyses consisted of obtaining means or scale values for each item. Correlations among methods were computed for each of the three attitude objects. Correlations for one well-researched attitude object were all above .80. Other correlations ranged from .41 to .89. Consistently the highest correlations were between rankings and paired comparisons, a finding that is not surprising in view of the fact that tasks required of these two scales are very similar.