Measure what you are trying to predict: Applying the correspondence principle to the Implicit Association Test

A Comparative Investigation of the Predictive Validity of Four Indirect Measures of Bias and Prejudice

Although measures of implicit associations are influential in the prejudice literature, comparative tests of the predictive power of these measures are lacking. A large-scale (N > 100,000) analysis of four commonly used measures-the Implicit Association Test (IAT), Single-Category IAT (SC-IAT), evaluative priming task (EPT), and sorting paired features task (SPF)-across 10 intergroup domains and 250 outcomes found clear evidence for the superiority of the SC-IAT in predictive and incremental predictive validity. Follow-up analyses suggested that the SC-IAT benefited from an exclusive focus on associations toward stigmatized group members, as associations toward non-stigmatized group members diluted the predictive strength of relative measures like the IAT, SPF, and EPT. These results highlight how conclusions about predictive validity can vary drastically depending on the measure selected and reveal novel insights about the value of different measures when focusing on predictive than convergent validity.

Trajectories of parents' gendered play attitudes during early childhood and implications for children's gender development

This study examined associations between parents' gendered attitudes about play and children's gender development. The sample was 501 families from a large US city followed annually for 4 years (501 mothers, 383 fathers; 69% White, 16% Latinx, 8% African American; children Mage  = 5.67 months, 53% boys). Latent trajectories examined change in parents' attitudes toward same- and other-gender play during first 4 years of the child's life. On a subsample (n = 252), trajectories were examined in relation to children's later gender-typed preferences and gender constancy. Parents grew more gender-flexible in their attitudes, particularly parents of boys. Trajectories reflecting more stereotypic attitudes showed some associations (small in magnitude) with children's gender-typed preferences and gender constancy by age 4.

Improving the predictor-criterion consistency of mindset measures: Application of the correspondence principle

Increasing the level of correspondence between measures of growth mindset and their related outcomes could afford more precise prediction of the relationships between growth mindset and social-emotional outcomes. To illustrate the value of measurement correspondence, two studies were conducted in Hong Kong. Study 1 showed that an agent-correspondent growth mindset measure (parents’ perception of the malleability of their children's personal qualities), compared to an agent-non-correspondent one (parents’ belief in the malleability of personal qualities of a generalized other), had stronger predictive relationship with children's likelihood of displaying difficult behaviors. Study 2 found that children's self-theories about the malleability of their intelligence (an intrapersonal construct) had stronger predictive relationship with academic engagement (an intrapersonal outcome) than did their perception of growth mindset norm (a normative construct). However, perceived growth mindset norm regarding personal qualities had stronger predictive relationship with peer relationship quality (an interpersonal outcome). Together these results demonstrated that when corresponding measures of growth mindset were used to predict an outcome, more reliable growth mindset effects would emerge.

Behavior prediction requires implicit measures of stimulus-goal discrepancies and expected utilities of behavior options rather than of attitudes toward objects

Meta-analyses show low correlations between implicit attitude measures and behavior measures, suggesting that these attitude measures are weak predictors of behavior. Researchers of implicit cognition have resorted to several rescue strategies. Their most important reply, based on a traditional dual-process theory of behavior causation, is that attitudes toward objects (positive/negative) automatically activate specific action tendencies (approach/avoidance), but that this stimulus-driven process can be overruled by a nonautomatic goal-directed process in which the expected utilities of action options are weighed up. According to such a theory, it makes sense to continue measuring attitudes with implicit measures, but research should also take into account the moderating role of goals and other factors. We propose an alternative dual-process theory in which goal-directed processes can also be automatic and count as the most important cause of behavior. According to this theory, the goal-directed process responsible for action selection is further preceded by the detection of a stimulus-goal discrepancy. Based on this alternative theory, we propose to no longer measure attitudes toward objects but rather to measure (a) the magnitude of stimulus-goal discrepancies as well as (b) the expected utility of the behavior at stake, understood as the product of the values of the outcomes of the behavior, and the behavior-outcome expectancies. Here too, implicit measures are needed because people may not always have conscious access to these constructs or be motivated to disclose them. This article is categorized under: Psychology > Theory and Methods Psychology > Emotion and Motivation Psychology > Reasoning and Decision Making.

The good and the bad: Are some attribute words better than others in the Implicit Association Test?

The Implicit Association Test (IAT) is one of the most popular measures in psychological research. A lack of standardization across IATs has resulted in significant variability among stimuli used by researchers, including the positive and negative words used in evaluative IATs. Does the variability in attribute words in evaluative IATs produce unwanted variability in measurement quality across studies? The present work investigated the effect of evaluative stimuli across three studies using 13 IATs and over 60,000 participants. The 64 positive and negative words that we tested provided similar measurement quality. Further, measurement was satisfactory even in IATs that used only category labels as stimuli. These results suggest that common sense is probably a sufficient method for selection of evaluative stimuli in the IAT. For reasonable measurement quality, we recommend that researchers using evaluative IATs in English select words randomly from the set we tested in the present research.

Predicting nonsuicidal self-injury using a variant of the implicit association test

OBJECTIVE

Nonsuicidal self-injury (NSSI) is a severe problem, and its prevalence is increasing. To aid prevention and treatment, there is an urgent need for evidence-based measures to identify individuals at risk for NSSI. Measures that probe past NSSI are most promising, but people are often motivated to conceal NSSI behavior. This problem can be overcome by using implicit measures, which do not require individuals to self-report on their behavior. Yet, prior research typically found weak predictive utility of implicit measures. Based on a new perspective on implicit measures and recent findings in NSSI research, we developed an Implicit Association Test that probes past NSSI (the P-NSSI-IAT).

METHOD

We report two preregistered studies (N = 83; N = 372) in which we tested the utility of the P-NSSI-IAT to detect past NSSI and predict NSSI one month later.

RESULTS

P-NSSI-IAT scores (a) differentiated injury groups from non-injury groups and (b) prospectively predicted NSSI and improved prediction above and beyond risk factors of NSSI.

CONCLUSIONS

These initial findings suggest that the P-NSSI-IAT is a promising tool for NSSI risk assessment. Future studies should further examine the predictive utility of this newly developed measure for NSSI behavior.

An investigation of emotional and evaluative implicit associations with police using four versions of the Implicit Association Test

In two studies (total N = 829), we assessed civilian implicit associations with police using four modified Implicit Association Tests (IAT). Across studies and IATs, individuals harbored stronger negative implicit associations (associating police with fear/bad) than positive implicit associations (associating police with safety/good). The predictive validity of the implicit associations and magnitude of D scores varied across IAT. In Study 1, the IATs involving categorization of police-related (vs. everyday) symbols were most sensitive, but the versions involving categorization of police (vs. civilian) models provided more evidence for predictive validity. In Study 2, the IAT involving categorization of emotional words (safety/fear) was most sensitive, but the version involving categorization of evaluative words (good/bad) provided more evidence for predictive validity. In both studies, we also assessed individual differences (race, political affiliation) in implicit associations. The findings prompt the need to further examine the underlying cognitive components of civilian attitudes toward police and emphasize the importance of developing several IATs when assessing implicit attitudes.

Specificity and incremental predictive validity of implicit attitudes: studies of a race-based phenotype

Four studies involving 2552 White American participants were conducted to investigate bias based on the race-based phenotype of hair texture. Specifically, we probed the existence and magnitude of bias in favor of Eurocentric (straight) over Afrocentric (curly) hair and its specificity in predicting responses to a legal decision involving the phenotype. Study 1 revealed an implicit preference, measured by an Implicit Association Test (IAT), favoring Eurocentric over Afrocentric hair texture among White Americans. This effect was not reducible to a Black/White implicit race attitude nor to mere perceptual preference favoring straight over curly hair. In Study 2, the phenotype (hair) IAT significantly and uniquely predicted expressions of support in response to an actual legal case that involved discrimination on the basis of Afrocentric hair texture. Beyond replicating this result, Studies 3 and 4 (the latter preregistered) provided further, and even more stringent, evidence for incremental predictive validity: in both studies, the phenotype IAT was associated with support for a Black plaintiff above and beyond the effects of two parallel explicit scales and, additionally, a race attitude IAT. Overall, these studies support the idea that race bias may be uniquely detected by examining implicit attitudes elicited by group-based phenotypicality, such as hair texture. Moreover, the present results inform theoretical investigations of the correspondence principle in the context of implicit social cognition: they suggest that tailoring IATs to index specific aspects of an attitude object (e.g., by decomposition of phenotypes) can improve prediction of intergroup behavior.

Information Matching: How Regulatory Focus Affects Information Preference and Information Choice

Individuals often prefer information that matches their needs. In this study, we aimed to explore the relationship between regulatory focus and information preference. Specifically, we investigated the effects of promotion-focused information and prevention-focused information on explicit and implicit information preferences and choice behavior, and examined the mediating roles of information preference. In Experiment 1, we found that prevention-focused individuals were more likely to choose functional information, whereas promotion-focused people were more likely to choose hedonic information. However, there was no significant relationship between regulatory focus and explicit preference and no mediating effect of explicit information preference. In Experiment 2, we found that promotion-focused individuals had a greater implicit preference for hedonic information than did prevention-focused individuals. Implicit information preference mediated the influence of regulatory focus on information choice. The findings of this study may help us understand the psychological mechanism underlying information preference and have important implications for information dissemination.

Relating Reactive and Proactive Aggression to Trait Driving Anger in Young and Adult Males: A Pilot Study Using Explicit and Implicit Measures

Driving anger and aggressive driving are main contributors to crashes, especially among young males. Trait driving anger is context-specific and unique from other forms of anger. It is necessary to understand the mechanisms of trait driving anger to develop targeted interventions. Although literature conceptually distinguished reactive and proactive aggression, this distinction is uncommon in driving research. Similar, cognitive biases related to driving anger, measured by a combination of explicit and implicit measures, received little attention. This pilot study related explicit and implicit measures associated with reactive and proactive aggression to trait driving anger, while considering age. The sample consisted of 42 male drivers. The implicit measures included a self-aggression association (i.e., Single-Target Implicit Association Test) and an attentional aggression bias (i.e., Emotional Stroop Task). Reactive aggression related positively with trait driving anger. Moreover, a self-aggression association negatively related to trait driving anger. Finally, an interaction effect for age suggested that only in young male drivers, higher proactive aggression related to lower trait driving anger. These preliminary results motivate further attention to the combination of explicit and implicit measures related to reactive and proactive aggression in trait driving anger research.

Moving Beyond the Relative Assessment of Implicit Biases: Navigating the Complexities of Absolute Measurement

A relative assessment of implicit biases is limited because it produces a combined summary evaluation of two attitudinal beliefs while concealing the biases driving this evaluation. Similar limitations occur for relative explicit measures. Here, we will discuss the benefits and weaknesses of using relative versus absolute (individual/separate) assessments of implicit and explicit attitudes. The Implicit Association Test (IAT) will be the focal implicit measure discussed, and we will present a new perspective challenging the evidence that the IAT can only be utilized to measure relative, not absolute, implicit attitudes. Modeling techniques (i.e., Quad models) that can determine the separate biases behind the relative summary evaluation will also be considered. Accurately utilizing absolute implicit bias scores will enable academia and industry to answer more complex research questions. For implicit social cognition to maintain and expand its usefulness, we encourage researchers to further test and refine the measurement of absolute implicit biases.

Reflecting on 25 Years of Research Using Implicit Measures: Recommendations for Their Future Use

For more than 25 years, implicit measures have shaped research, theorizing, and intervention in psychological science. During this period, the development and deployment of implicit measures have been predicated on a number of theoretical, methodological, and applied assumptions. Yet these assumptions are frequently violated and rarely met. As a result, the merit of research using implicit measures has increasingly been cast into doubt. In this article, we argue that future implicit measures research could benefit from adherence to four guidelines based on a functional approach wherein performance on implicit measures is described and analyzed as behavior emitted under specific conditions and captured in a specific measurement context. We unpack this approach and highlight recent work illustrating both its theoretical and practical value.

Twenty-Five Years of Research Using Implicit Measures

The year 2020 marks the 25th anniversary of two seminal publications that have set the foundation for an exponentially growing body of research using implicit measures: Fazio, Jackson, Dunton, and Williams's (1995) work using evaluative priming to measure racial attitudes, and Greenwald and Banaji's (1995) review of implicit social cognition research that served as the basis for the development of the Implicit Association Test (IAT). The current article provides an overview of (1) two conceptual roots that continue to shape interpretations of implicit measures; (2) conflicting interpretations of the term implicit; (3) different kinds of dissociations between implicit and explicit measures; (4) theoretical developments inspired by these dissociations; and (5) research that used implicit measures to address domain-specific and applied questions. We conclude with a discussion of challenges and open questions that remain to be addressed, offering guidance for the next generation of research using implicit measures.

Decomposing implicit associations about life and death improves our understanding of suicidal behavior

The Death/Suicide Implicit Association Test (IAT) is effective at detecting and prospectively predicting suicidal thoughts and behaviors. However, traditional IAT scoring procedures used in all prior studies (i.e., D-scores) provide an aggregate score that is inherently relative, obfuscating the separate associations (i.e., "Me = Death/Suicide," "Me = Life") that might be most relevant for understanding suicide-related implicit cognition. Here, we decompose the D-scores and validate a new analytic technique called the Decomposed D-scores ("DD-scores") that creates separate scores for each category ("Me," "Not Me") in the IAT. Across large online volunteer samples (N > 12,000), results consistently showed that a weakened association between "Me = Life" is more strongly predictive of having a history of suicidal attempts than is a stronger association between "Me = Death/Suicide." These findings replicated across three different versions of the IAT and were observed when calculated using both reaction times and error rates. However, among those who previously attempted suicide, a strengthened association between "Me = Death" is more strongly predictive of the recency of a suicide attempt. These results suggest that decomposing traditional IAT D-scores can offer new insights into the mental associations that may underlie clinical phenomena and may help to improve the prediction, and ultimately the prevention, of these clinical outcomes.

Can the Implicit Association Test Serve as a Valid Measure of Automatic Cognition? A Response to Schimmack (2021)

Much of human thought, feeling, and behavior unfolds automatically. Indirect measures of cognition capture such processes by observing responding under corresponding conditions (e.g., lack of intention or control). The Implicit Association Test (IAT) is one such measure. The IAT indexes the strength of association between categories such as “planes” and “trains” and attributes such as “fast” and “slow” by comparing response latencies across two sorting tasks (planes–fast/trains–slow vs. trains–fast/planes–slow). Relying on a reanalysis of multitrait–multimethod (MTMM) studies, Schimmack (this issue, p. 396) argues that the IAT and direct measures of cognition, for example, Likert scales, can serve as indicators of the same latent construct, thereby purportedly undermining the validity of the IAT as a measure of individual differences in automatic cognition. Here we note the compatibility of Schimmack’s empirical findings with a range of existing theoretical perspectives and the importance of considering evidence beyond MTMM approaches to establishing construct validity. Depending on the nature of the study, different standards of validity may apply to each use of the IAT; however, the evidence presented by Schimmack is easily reconcilable with the potential of the IAT to serve as a valid measure of automatic processes in human cognition, including in individual-difference contexts.