lambert

Attitudes as Dominant Responses—Why Public Settings Can Exacerbate Racial Prejudice : 

Attitudes as Dominant Responses—Why Public Settings Can Exacerbate Racial Prejudice Alan Lambert Washington University Collaborators: Keith Payne Larry Jacoby Lara Shaffer Alison Chasteen Saera Khan

Today’s Talk: 

Today’s Talk Brief literature review Attitudes as Dominant Response (ADR) Model   3 Experiments Experiment 1: Impression formation Experiment 2: Reaction time Experiment 3: Stereotypic errors in weapon identification Unresolved issues/ongoing research

Experimental investigations of Private vs. Public Contexts: 

Experimental investigations of Private vs. Public Contexts Common in several domains (accountability, impression management, conformity, attitude change) Surprisingly understudied in stereotyping area Blanchard et al, 1991; Dutton & Yee, 1974; Lambert et al. 1996; Monteith et al, 1996; Plant & Devine, 1998 Mixed implications

Attitudes as Dominant Response ModelLambert, Payne, Shaffer, Jacoby, Chasteen, & Khan (under review) : 

Attitudes as Dominant Response Model Lambert, Payne, Shaffer, Jacoby, Chasteen, & Khan (under review)   Intuitive assumptions regarding stereotyping and public contexts may not be correct Attempt to bridge two lines of research: Impact of actual/imagined presence of others on task performance (e.g. Triplett, 1898) The literature on “attitude-behavior consistency” (e.g. LaPiere, 1934)

Very brief overview of social facilitation literature: 

Very brief overview of social facilitation literature Is performance improved or impaired in “public” (audience or co-actor) conditions ? Zajonc (1965; see also Hull, 1943): Habitual/dominant responses more likely in public  Resolution: If dominant response yields correct answer: helps performance If dominant response yields incorrect answer: hurts performance

Why would public settings make dominant responses more likely?: 

Why would public settings make dominant responses more likely? Audience Generalized arousal/anxiety Facilitation of dominant responses Drive/arousal is an “intensifier” in that it “…adds fuel to whatever fire is burning at the time” (Allen et al. 1989) ? much debate as to exact reason

Important class of mental habits: attitudes : 

Important class of mental habits: attitudes Attitude object (S)  evaluative reaction (R) Idea of mental habits is not new (James, 1890) but… We believe that we are the first (?) to make an explicit connection between the social facilitation literature and current research/theory on attitude activation and application

Implications of ADR model: 

Implications of ADR model Extremely counterintuitive prediction: If stereotypic attitudes are mental habits, then: use of these attitudes should be greater in public compared to private, especially among participants high in social anxiety Thus: stereotype use could be higher in precisely the situation in which you’d think it’d be most unlikely! Evaluative response private

Implications of ADR model: 

Implications of ADR model Extremely counterintuitive prediction: If stereotypic attitudes are mental habits, then: use of these attitudes should be greater in public compared to private, especially among participants high in social anxiety Thus: stereotype use could be higher in precisely the situation in which you’d think it’d be most unlikely! Evaluative response public

Experiment 1: 

Experiment 1 Racial attitudes pre-measured two months earlier Exact way that sentiments are measured doesn’t matter (e.g. modern racism vs. social dominance) Impression formation task Subtly identified as Black in all cases Ambiguous individuating information presented Dependent variable: evaluative and trait ratings Participants complete impressions under one of two sets: private vs. anticipated public Post-test: Trait differences in social anxiety measured

Slide11: 

Part one: Impression Formation In this part, we will be asking you to form an impression of another individual….. Part Two: Public Discussion In real life, we often share our judgments with other people. Therefore, after you have expressed your judgments, there will be a general discussion session with the other participants in the study today. During this discussion, you will be asked to share and discuss your judgments with the other participants, who also participated in this task today… Sample of anticipated public instructions

Slide12: 

Experiment 1 Attitude-behavior consistency Hypothesis “Habit strengthening” Cognitive Load Reduction of Control stronger stronger stronger Predictions for Public Setting (compared to Private)

Slide13: 

Results: Relation between racial attitudes and judgments of target Low anxiety participants High anxiety participants

Slide14: 

Experiment 1 Attitude-behavior consistency Hypothesis “Habit strengthening” Cognitive Load Reduction of Control stronger stronger stronger Predictions for Public Setting (compared to Private)

Experiment 2: 

Experiment 2 Participants complete Fazio-type attitude RT task for a series of 30 attitude objects spanning large range of topics (e.g. affirmative action, legalization of marijuana, Al Gore, gun control, etc.) Main DV: response latency to make “good” or “bad” responses

Slide16: 

Experiment 1 Consistency of impressions with stereotypic attitudes Experiment 2 Reaction time Hypothesis “Habit strengthening” Cognitive Load Reduction of Control stronger stronger stronger faster faster slower Predictions for Public Setting (compared to Private)

Slide17: 

Presentation of attitude object Automatic processes (Fast and effortless) Controlled processes (Slow and effortful) Physical response Factors that can reduce controlled processing: e.g. response deadlines, motivation and (we believe) public contexts in presence of arousal/anxiety Reduction of Control Hypothesis

Results: Experiment 2: 

Results: Experiment 2 Private Anticipated Public Difference High trait anxiety Low trait anxiety 1975 1626 1381 1560 -349 +179 Regression analyses: F (1,46) = 4.37, p < .05 for 2-way Context X Anxiety interaction

Slide19: 

Experiment 1 Attitude-behavior consistency Experiment 2 Reaction time Hypothesis “Habit strengthening” Cognitive Load Reduction of Control stronger stronger stronger faster faster slower Predictions for Public Setting (compared to Private)

Experiment 3:Stereotypic errors in weapon identification: 

Experiment 3: Stereotypic errors in weapon identification Based on paradigm used by Payne (2001, JPSP) Amidou Diallo case Congruent black prime/threatening target white prime/non-threatening target Incongruent black prime/non-threatening target white prime/threatening target Goal of our study: Demonstrate generalizability of our counterintuitive findings Leverage in teasing apart reason WHY using Jacoby’s (1991) process dissociation procedure

Slide21: 

500 ms 200 ms 100 ms 550 ms deadline Design: Prime (Black vs. White) Target (gun vs. tool) Context (Private vs. Anticipated public) DV: Identification of object as gun vs. tool

Slide22: 

Experiment 1 Attitude behavior consistency Experiment 2 Reaction time Experiment 3 Hypothesis “Habit strengthening” Cognitive Load Reduction of Control stronger stronger stronger faster faster slower Predictions for Public Setting (compared to Private) Stereotypic Errors more more more

Slide23: 

Private Public Proportion of Errors Black face White face Race X Target X Context p < .01

Slide24: 

Experiment 1 Attitude-behavior consistency Experiment 2 Reaction time Hypothesis “Habit strengthening” Cognitive Load Reduction of Control stronger stronger stronger faster faster slower Predictions for Public Setting (compared to Private) Stereotypic Errors more more more Experiment 3

Slide25: 

In this paradigm, participants have the goal of (a) responding “gun” when the target is, in fact, a gun, and (b) of responding “tool” only when the target was actually a tool. According to PDP, control is operationalized as the ability to flexibly monitor and control one’s responses, therefore to successfully discriminate between guns and lures. What happens when control fails: PDP assumes that participants use an alternate basis of responding, based on the most accessible knowledge at the time. (race of prime exerts effect here) PDP assumes that automatic and controlled processes are two independent bases for responding (cf. Jacoby et al. 1997; Hinztman & Curran, 1997) Thumbnail sketch of process dissociation assumptions

Slide26: 

Experiment 1 Attitude-behavior consistency Experiment 2 Reaction time Hypothesis “Habit strengthening” Cognitive Load Reduction of Control stronger stronger stronger faster faster slower Predictions for Public Setting (compared to Private) Stereotypic Errors more more more Experiment 3 Process dissociation estimates Cognitive control Accessibility bias no change no change no change higher lower lower

Slide27: 

Cognitive Control Estimates Prime Race Black White Public Private .44 .45 .54 .51 Effect of context p < .05; no effect of prime race

Slide28: 

Accessibility bias Estimates probability to respond “gun” when control fails Prime Race Black White Public Private .38 .33 .41 .35 prime race p < .001; no effect of context

Slide29: 

Note double dissociation, consistent with independence assumption: prime affects accessibility bias, but not control. context affects control, but not accessibility bias.

Slide30: 

Experiment 1 Attitude-behavior consistency Experiment 2 Reaction time Hypothesis “Habit strengthening” Cognitive Load Reduction of Control stronger stronger stronger faster faster slower Predictions for Public Setting (compared to Private) Stereotypic Errors more more more Experiment 3 Process dissociation estimates Cognitive control Accessibility bias no change no change no change higher lower lower

Role of Anxiety: 

Role of Anxiety Further analyses show that effects of context on control are moderated by anxiety (but complicated). Translation: I couldn’t finish analyses prior to SESP

Summary: 

Summary Across three experiments: greater stereotyping in public compared to private, primarily among high anxiety participants Tested the viability of three process-level explanations (habit strengthening, cognitive load, reduction of control) Our results speak more generally to the social facilitation literature. Theorists have long debated the inability of the cognitive load explanation to fully explain social facilitation effects The reduction of control hypothesis may provide a more viable explanation (?)

Caveats and directions for future research: 

Caveats and directions for future research Explore different kinds of public contexts Is there something special about anxiety? Moderation vs. mediation In the stereotyping area: Further integrate the “cognitive load” and social facilitation literatures

Slide35: 

Congruent = probability of responding gun on a congruent trial (B prime  gun) C + A (1-C) C = control A = accessibility bias Incongruent = probability of responding gun on a incongruent trial (B prime  tool) A (1-C) Solving: Estimates of C = congruent – incongruent Estimates of A = incongruent/1-C