Extending the Implicit Association Test (IAT): Assessing Consumer Attitudes Based on Multi-Dimensional Implicit Associations

Carbon C-C (2011) Extending the Implicit Association Test (IAT): Assessing Consumer Attitudes Based on Multi-Dimensional Implicit Associations. PLoS ONE 6(1): e15849. doi:10.1371/journal.pone.0015849 Extending the Implicit Association Test (IAT): Assessing Consumer Attitudes Based on Multi-Dimensional Implicit Associations Valentin Gattol 0 Maria Sa a ksja rvi 0 Claus-Christian Carbon 0 Susanne Hempel, Rand, United States of America 0 1 Department of Product Innovation Management, Faculty of Industrial Design Engineering, Delft University of Technology , Delft , The Netherlands , 2 Department of General Psychology and Methodology, University of Bamberg , Bamberg , Germany Background: The authors present a procedural extension of the popular Implicit Association Test (IAT; [1]) that allows for indirect measurement of attitudes on multiple dimensions (e.g., safe-unsafe; young-old; innovative-conventional, etc.) rather than on a single evaluative dimension only (e.g., good-bad). Methodology/Principal Findings: In two within-subjects studies, attitudes toward three automobile brands were measured on six attribute dimensions. Emphasis was placed on evaluating the methodological appropriateness of the new procedure, providing strong evidence for its reliability, validity, and sensitivity. Conclusions/Significance: This new procedure yields detailed information on the multifaceted nature of brand associations that can add up to a more abstract overall attitude. Just as the IAT, its multi-dimensional extension/application (dubbed mdIAT) is suited for reliably measuring attitudes consumers may not be consciously aware of, able to express, or willing to share with the researcher [2,3]. - Traditionally, attitudes have been measured by having consumers respond to an attitude object (or entity) on self-report rating scales. In these scales, consumers rate a particular object (e.g., a product or a brand) on dimensions such as good/bad, like/dislike, or pleasant/unpleasant. Yet, consumers often find it difficult to report on these scales. They may not have attitudes readily available for reporting on them (in an explicit way), or may even find it difficult to retrieve them [2,3]. Indirect measures, in particular the popular Implicit Association Test (IAT) by Greenwald, McGhee, and Schwartz [1], constitute a viable alternative avoiding some of the problems associated with direct measures (e.g., lack of attitude availability/accessibility, social desirability bias). In this article, we introduce a procedural extension of the IAT, a multi-dimensional Implicit Association Test (md-IAT). In contrast to the regular IAT, which is utilized as a procedure that allows assessment on a single dimension only, the md-IAT comprises six dimensions, thus allowing for a more detailed, multi-dimensional assessment of attitudes. More finegrained attitudes/associations have been assessed in several studies but were confined to a single administration and thereby also to a single dimension in the IAT: for example, to measure gender stereotypes (i.e., menwomen/warmcold [4]), self-concepts (i.e., selfother/anxiouscalm, [5]), or even abnormal pedophilic tendencies (i.e., childrenadults/sexno-sex, [6]). The additional information offered by this multi-dimensional measure can be of particular value in marketing and consumer research, allowing for examplein the same way as with direct measuresto easily create more complex and differentiated profiles of products and brands (cf. [7]). Tapping consumer insights in such a way more appropriately captures the richness of consumers perceptions, feelings, and attitudes toward a brand. For example, the IAT can indeed provide important information about consumers general attitude toward a specific brand or product (consumers likes and dislikes), but it does not elucidate the different components contributing to this global attitude. Any kind of intervention, however, depends on clear diagnostics: the specific aspects consumers like or dislike or the specific properties they associate with the product [8]. The contribution of the present research is both of theoretical and practical relevance: our results show that the md-IAT procedure is a methodologically sound extension of the IAT thatunlike the latteralso allows for multi-dimensional assessment of brand attitudes. This in turn opens up numerous possibilities for researchers to test constructs such as brand or product personality [9,10], or more generally, consumers brand associations or attitudes on any kind of multi-dimensional scale [11]. In addition, we show that the md-IAT, just like the IAT, is not affected by the specific stimuli selected to represent a brand. The three brand identifiers used in the present studies (logos, signatures, and product pictures) all yielded similar results, therefore rendering the md-IAT rather suited as a conceptual (as opposed to perceptual) measure of brand attitudes. The structure of the paper is as follows: We start by briefly reviewing different forms of attitude measurementdistinguishing between indirect and direct measures. We then turn to the IAT itself before introducing its multi-dimensional extension (the md-IAT) and its application in two within-subjects repeated measurement studies. Indirect versus Direct Measures Indirect measures differ from direct measures in that they do not rely on verbal self-reports as a way of inferring attitudes [12]. Instead, they rely on rather indirect means of assessing an attitude, for example differences in reaction times, facial expression, or specific brain activation. Indirect measures can be further distinguished into physiological or latency based measures. Physiological measures include techniques such as electro-dermal activity (EDA; [13]), pupillometry [14], eyetracking [15], electromyography (EMG; [16]); or various brain imaging techniques, such as functional magnetic resonance imagining (fMRI; [17]), which allow direct observation of brain activity during mental tasks. While promising in their own right, these physiological measures do not yet offer standardized forms of attitude assessment (for advances in this domain, see, [16,18]). In addition, they require (very) expensive equipment and a considerable expertise in the domain of cognitive neuroscience, which make most of these research techniques inaccessible and/or ill-suited for any kind of more applied research. This is much less the case for indirect measures based on response latencies (or reaction times). Measures such as affective priming [19], the Extrinsic Affective Simon Task [20], the Go/No-Go Association Task [21], and particularly the Implicit Association Test (IAT; [1]), are fairly standardized forms of attitude assessment requiring little more than a computer and a testing environment void of external distractions. Attitude Measurement and the Implicit Association Test (IAT) The IAT is a method of estimating evaluative associations that underlie im (...truncated)