RESURRECTING THE ERROR CHOICE TECHNIQUE: The premature demise of an indirect measure of attitude?
Porter, Ronald D.
MetadataShow full item record
The error choice (EC) technique was among the earliest indirect attitude measures developed and was originally designed to overcome social desirability concerns (Hammond, 1948). This programme of research set out to advance EC research in several ways. First, an exploratory factor analysis examined whether participants’ responses to the EC target items produced a systematic pattern of responding. The factor analysis indicated that a single underlying factor best accounted for the data. Additionally, the EC measure demonstrated good reliability across the 3 studies. Second, these studies provided evidence that the EC measure is, at least in part, attitudinal. The EC measure showed a modest positive correlation with the direct measure of attitude in all 3 studies. This suggested that participants’ responses to the EC target items were, at least partially, attitudinal. Additionally, across the studies participant’s EC scores did not change between the high and low social desirability conditions, while participants’ scores on the direct measure were significantly more positive in the high social desirability condition. These findings suggest that the EC measure is, to some degree, resistant to socially desirable responding. Studies 2 and 3 also represent the first time that recommendations made by early EC researchers to improve the EC technique were systematically examined. In these studies the amount of time participants had to complete the EC measure was restricted. The time restriction did not improve the performance of the EC measure. The other optimal condition examined in Study 3 was the presence and absence of filler items in the EC measure. Indeed, removing filler items from the EC measure did not negatively impact its performance. Finally, this programme of research compared the EC measure with more contemporary indirect measures of attitude. In Study 2, the Implicit Association Test (IAT) was compared with the EC and direct measure. In Study 3, the Personalized IAT was compared to the EC and direct measure. The results revealed that neither the IAT nor PIAT correlated with the EC measure. In summary, these results suggest the EC technique holds some promise as an approach to attitude measurement and is well worth resurrecting.