Methodological, Social and Clinical Applications of the Implicit Relational Assessment Procedure (IRAP)

The aims of the current thesis were three fold. The first aim was primarily methodological and attempted to determine the stimulus presentation that would produce the strongest IRAP effect. Experiment 1 set out to address this simple experimeental question. The key manipulation involved the presence or absence of the randomisation of the sample stimuli and/or the respponse options within the IRAP. The results indicated that the ramdomisation of the sample stimuli (with or without the randomisation of the response options) generated the strongest D-IRAP scores. Randomising the response options alone without randomising the samples also produced a significant IRAP effect, although this was smaller. The D-IRAP score that resulted from the static presentation of both samples and response options was almost negligible. The second aim of the thesis was to determine the utility of the IRAP as a measure of implicit attitudes to race and this was the core focus of Experiments 2 to 6. In the IRAP in Experiment 2, participants were simply required to relate the phrases WHITE PERSON and BLACK PERSON to a range of simple positive and negative terms that could be readily categorised as safe and dangerous. We predicted that our Irish participants would show a pro- White and anti- Black bias. The D-IRAP trial-type scores were in part consistent with experimental predictions and showed a pro- White bias where White was safe, but not an anti-Black bias. In fact, participants significantly confirmed that Black was safe. This coincided with the explicit measures, on which participants indicated the absence of racial bias against Black people. In line with existing IAT evidence, to determine whether racial pictures would generate stronger IRAP effects than words, and to examine the potential role of attentional weapon bias, the IRAP in Experiment 3 presented pictures of Black and White men holding guns. For correct responding, however, the presence of the guns was incidental and the discriminations among the pictures were based on race alone. We predicted that Irish participants would show a pro- White and anti- Black bias that was perhaps enhanced by the presence of the guns. Participants significantly confirmed that White was safe and Black was dangerous, although they also significantly confirmed that Black as safe. Hence, this was our first evidence of an anti- Black bias, but other effects were not entirely as expected. Again on the majority of explicit measures, participants indicated the absence of racial bias. In order to determine whether the anti-Black bias in the previous study did reflect the influence of the guns, Experiment 4 presented pictures of Black and White men holding mundane objects. We predicted that Irish participants would show a pro- White and anti- Black bias, even though both were holding mundane objects. Participants significantly confirmed that both White and Black were safe, but had mixed views of either as dangerous. Hence, the anti- Black bias from the previous study was now lost, suggesting that the weapons had facilitated the effect. Again, on the majority of the explicit measures, participants indicated the absence of racial bias. In line with previous IAT studies of race and to examine further the outcomes from the previous studies, Experiment 5 presented pictures of Black and White men holding both guns and mundane objects. We predicted that Irish participants would show a pro- White and anti- Black bias that may be differentially influenced by the items presented in the pictures. The four trial-types examined in the previous studies were sub-divided into eight to permit comparisons of the relative influence of the different items held in the pictures. Overall, seven of the eight trial-types were nonsignificant and participants only significantly confirmed that White men with mundane objects were safe. In fact, close inspection of the non-significant effects indicated unexpected evidence of pro- Black bias. Again these agreed with the explicit measures, on which participants indicated the absence of racial bias. In Experiment 6, participants were presented with the same IRAP as the previous study, but were required to discriminate the objects in the pictures as the basis for correct responding and ignore race. That is, on consistent trials guns were always dangerous and mundane objects were always safe and correct responding on inconsistent trials was reverse (guns always safe and mundane objects dangerous). We predicted that Irish participants would categorise guns as dangerous and mundane objects as safe, but expected that these effects might be influenced by race. In this study, six of the eight were non-significant. Participants only significantly confirmed that Black with a gun was not safe and White with a gun was dangerous, but there was no evidence of clear influence of either items or race in the remaining effects. Again, the explicit measures indicated the absence of racial bias. The third aim of the current thesis was to examine the utility of the IRAP as a measure of clinically relevant implicit cognitions. Previous preliminary research had indicated strong IRAP effects that showed a positive bias towards the self in participants with a high explicit self-esteem. In Experiment 7, participants were required to indicate their agreement or disagreement with a range of statements that reflected high self-esteem (i.e. were positive about the self) or reflected low selfesteem (i.e. were negative about the self). Because the study was aslo concerned with whether on not the IRAP would correlate with an explicit measure of self, the positive and negative statements presented in the IRAP were taken directly from the Roseberg Self-Esteem Scale (RSES) and particiants also completed this as an explicit measure. We predicted that participants high on explicit self-esteem would relate the self with the positive self-statements and not relate the self with negative self-statements. The findings indicated that participants significantly confirmed that they agreed with the positive (high self-esteem) statements and significantly disconfirmed that they disagreed with them. They also disconfirmed that they agreed with the negaative statements, although this effect was non-significant. Although these findings were largely as expected from participants who all scored as high in selfesteem on the explicit RSES, there was no significant correlation between the two measures. In Experiment 8, we used the IRAP to explore implicit attitudes towards acceptance and avoidance. For this purpose, we selected undergraduate students who showed an xplicit propensity towards high acceptance/ low avoidance on the Acceptaance and Action Questionnaire (AAQ-2). In order to enhance the potential overlap between the two types of measure, as we had done in Experiment 7, we took opposing satements directly from the AAQ and inserted them into the IRAP as target stimuli. The results indicated that participants significantly confirmed that they agreed with the acceptance statements, but all other effects were non-significant. Closer inspection of the trial-types, however, were as expected and showed that particpants disconfirmed that they agreed with the avoidance statements and confirmed that they disagreed with them. In spite of these outcomes, the IRAP data did not correlate significantly with the AAQ. The current reaearch program offered preliminary investigations of the utility of the IRAP as a measure of impicit attitudes to a range of psychological phenomena, including race, self-esteem and acceptance. Although the IRAP effects in many cases were not as significant and predicted, particularly in the context of race, there wasgood evidence thata further investigation would be promising. The findings here make a useful contribution to the small existing IRAP research base by highlighting key issues that influence IRAP outcomes in a range of areas