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 Objects7pressing thespacekey,having memorizedthetargetpicture and its converted valence.Each main blockconsisted of 32 trials. Since half were identical tothe 16 trials in the practice block, the first pictureand second picture werefiller pictures that differedfrom each other (filler trials). In contrast, on theother 16trials,oneofthefiller pictures appeared asthefirst pictureand thetarget picturefor theblockappeared as the second picture (target trials). Ontarget trials,each of the four filler pictures appearedas the first picture four times. On each mainblock, the presentation of the 32 main trials wasrandomized for each participant. The participant’stask was identical to the practice block of theSECT,with theexception ofconverting theoriginalvalence of the target picture according to theinstructions for theblock.Therewere20-sintervalsbetween main trials. Participants completed all sixmain blocks according to the instructions thatappeared on the monitor.- .After completing the twoindirect measures, participants rated the pleasantnessofthreetarget pictures and four filler pictures,which appeared on the screen individually in randomorder.Participants rated each pictureonanine-point scale ranging from 1(unpleasant)to 9 (pleasant)andpressed the corresponding numerical key.The entire experimental procedure took approximately30 min.ResultsFUMIE test.Erroneous responses wereexcludedfrom the analysis (3.80% of all observations).Additionally,responseswith short latencies ofbelow300msandlong latenciesofover1500mswereexcludedas outliers (0.74% of all observations). Table 2shows the mean response latency and number oferrors for each of the six main blocks. For eachparticipant,threeattitudescores pertaining to threetarget pictures were obtained by subtracting themean response latency for the positive block fromthe mean response latency for the negative block(Table 2). Magnitude relationships for the threeattitude scores obtained were matched with theprediction. However, a one-way within-participantsANOVAwithattitudescoresasthedependentvariablerevealed that therewas no significant maineffect (F(2, 24)＝3.20, n.s,η?＝.17). This indicatesthat the FUMIE test did not differentiate significantlybetween the three different target pictures.In order to investigate the split-half reliabilityofthe FUMIE test, 26 trials included in each mainblock were divided evenlyinto two parts.For eachparticipant, two attitude scores were obtained foreach target picture, the correlation coefficient forthose two attitude scores was calculated, and anaverage of the three correlation coefficients taken.Finally,as an adjusted reliabilitycoefficient,ρ＝.83was obtained byapplying Spearman-Brown correctionto the average of the three correlation coefficients.SECT. Erroneous responses were excluded fromthe analysis (8.49% of all observations). Additionally,responseswith short latencies ofbelow 300msand long latencies ofover 1500ms wereexcluded asoutliers (3.21% of all observations). Table 3 showsTable 2Average Response Latencies (ms), Number of Errors, and Attitude Scores for FUMIE in Experiment 1Block123456Original valence oftargetPositivePositiveNegativeNeutralPositiveNegativeNegativePositiveNegativeConverted valenceoftargetResponse latency Number of errors499 (72)564(115)514 (77)524 (81)545 (71)519(125).92 (.95)1.15(1.28).77 (.83).92(1.12)1.38(1.80).77 (.93)Note.Attitude score represents the difference in average response latencies between two blocks.Digits in parentheses are standard deviations.Attitude score fortarget6510-26