ブックタイトル明星大学 心理学年報 第32号

概要

明星大学 心理学年報 第32号

Hayashi：Development and Testing of a New Indirect Attitude Measure for Pictorial Objects9Figure 2. Scatter plots of participants’pleasantness ratings for target pictures (y-axis) and indirectlymeasured attitude scores of target pictures (x-axis) in Experiment 1. The independent variable(x-axis) of the left diagram is the attitude score obtained by the FUMIE test. The independentvariable (x-axis) of the right diagram is the attitude score obtained by the SECT.standard deviation of participants’ratings for eachtarget picture was extremelysmall(see Table 4).Itis therefore inferred that their ratings tended toreflect knowledge common to their culture,?ratherthan participants’individual variability. In situationsin which a variable does not vary,calculatinga correlation coefficient between variables isthought to be unproductive.DiscussionAnalysis indicated that the FUMIE test was notan effective attitude measure for pictures, whereasthe SECT had better predictive power. Results oftheANOVA on theattitudescores from theFUMIEtest indicated that theFUMIE test did not discriminateat all between three emotionallydifferent targetpictures.Conversely,three attitude scores fromthe SECT were significantly different with respectto two ofthreecomparison pairs.Simpleregressionanalysis clearly indicated that the SECT was thesuperior predictor for the self-report measure.Although the reliability coefficient for the SECTwaslowerthanfortheFUMIE test,thereliabilityoftheSECT was almost comparableto theIAT,showingexceptionallyhigh reliabilityscores for attitudemeasures using response latencies.It has been reported that indirect attitudeand self-report measures do not often correlate significant-? For example,theproposition“An unclean lavatorybasinis unpleasant”would be common knowledgeamong peopleliving in developed countries.ly (for a review, see Dijksterhuis, Albers, & Bongers,2009).Asan exampleofa lower correlation,ina meta-analysis conducted by Dasgupta, McGhee,Greenwald, and Banaji (2000), r＝.12 was reportedas the overall correlation between self-report andattitude scores obtained via the IAT.Although it isdifficult to directly compare the R squared calculatedin the present study with correlation coefficientsreported in previous research,it is possibletoconclude that attitude scores from the SECT werebetterpredictorsoftheself-report measure,relativetotheFUMIE test,andtheSECT’spredictivepowerwas satisfactory.In conclusion, Experiment 1 suggested that theSECT had better potential than the FUMIE test asan indirect attitude measure for pictures. Experiment2 further investigated the efficiency of theSECT.Experiment 2One methodological feature of the SECT is theuse of positive and negative filler pictures asattitudinal objects along with target pictures. AsExperiment 1 fixed filler pictures for all participants,it is possible that the SECT attitude scoresobtained in Experiment 1mayhavebeen an artifactderived from the specific filler pictures used in theexperiment.That is,ifattitudescores changesignificantlyaccording to the filler pictures, it could beconcluded that the practical utility of the SECT is