**Participants** Participants will be non-Black individuals drawn from one of three sources: Mechanical Turk, students in a large undergraduate psychology course at Brock University, or participants at the research website Project Implicit (http://implicit.harvard.edu) who have not completed an IAT on Project Implicit before. **Procedure** Procedurally, the experiment will be similar to Study 4 in Lai et al., 2013. Half of the participants will be randomly assigned to take a pretest Race IAT and half of the participants will be assigned to take nothing at all. Participants will then be randomly assigned to complete one of nine effective manipulations from Study 3 of Lai et al., 2013, or a control condition in which no manipulation is administered. After, they will take a Race IAT and a measure of explicit racial attitudes. Participants will be invited to participate in a follow-up session. Between two and four days after they complete the first session (and with a reminder after another 2 days), consenting participants will be emailed to participate in a second session where they will complete the Race IAT and a measure of explicit racial attitudes again. They will also complete two social judgment items that assess support for pro-Black affirmative action and an item assessing their effort in the study. Based on prior multiple-session studies conducted online (Ranganath & Nosek, 2008), we anticipate that the mean length of time between initial and follow-up sessions will be about 5 days. We anticipate moderate attrition in the Project Implicit and MTurk samples, but little attrition in the psychology course sample. **Dependent Measures** **Implicit Association Test.** The IAT assesses the relative strength of associations between two conceptual categories (i.e., White people, Black people) and two evaluative attributes (i.e., Good, Bad; see Nosek, Greenwald, & Banaji, 2007, for a review). The procedure will follow the recommendations by Nosek, Greenwald, and Banaji (2005). Participants will be instructed to categorize words and images as quickly as possible while also being accurate. The IAT iwe used s composed of five blocks, with three practice blocks (omitted for analyses) and two critical blocks. In the first practice block (16 trials), participants categorize images of Black faces and White faces to categories labeled on the left or right. In the second practice block (16 trials), participants categorize good and bad words. In the third (32 trials) critical block, participants categorize images of Black faces/White faces and good/bad words on alternating trials. Consequently, participants categorize Black faces and bad words with one key and White faces and good words with another key. In the fourth practice block (24 trials), participants categorize images of White and Black faces again, except the White face/Black face categories have switched sides. The face category originally on the left is now categorized with the right key, and the face category originally on the right is now categorized with the left key. In the fifth (32 trials) critical block, participants categorize pairings opposite to the ones in the third block. Consequently, participants categorize White faces and bad words with one key and Black faces and good words with the other key. The fifth block will be counterbalanced with the third block between participants to control for potential order effects (Greenwald et al., 1998). The background of the webpage (i.e., Black or White), stimulus sets (one of two sets of image/word stimuli), and the evaluative attribute categories (i.e., Good/Bad or Pleasant/Unpleasant) were randomized for each IAT. Half of the time, participants will receive an IAT with a Black background, one set of image/word stimuli, and Good/Bad as evaluative categories. The other half of the time, participants will receive an IAT witha White background, the other set of image/word stimuli, and Pleasant/Unpleasant as evaluative categories. This randomization was done to control for potential testing effects related to specific stimuli or contexts. The IAT will be scored with the D algorithm recommended by Greenwald, Nosek, and Banaji (2003). A positive D score indicates faster responding on average when White faces were paired with good words and Black faces were paired with bad words compared to the reverse. Positive scores are interpreted as an implicit preference for White people compared to Black people. **Self-reported racial attitudes.** Participants will complete three self-report items measuring racial attitudes. One will assess relative preference for White people over Black people on a seven-point Likert scale ranging from “I strongly prefer Black people to White people” to “I strongly prefer White people to Black people.” The others will be feeling thermometers for White people and Black people measured using a seven-point scale ranging from “Very cold” to “Very warm.” For analyses, a difference score will be computed between the two feeling thermometers and averaged with the racial preference measure after standardizing each (SD = 1) while retaining a rational zero point of no preference between White people and Black people (α = .69, .69, .70). More positive scores will indicate a greater explicit preference for White people over Black people. **Social judgment.** Participants will complete two self-report items measuring judgments related to affirmative action in the second session. Both items have a "Yes"/"No" as response options. The items were: "A corporate personnel officer is evaluating a Black job applicant and a White job applicant who are identically qualified except the White applicant has more prior experience in related work. Is there a reasonable justification for this personnel officer hiring the Black applicant rather than the White applicant?" and "A college admissions officer considers applications from Black applicants and White applicants with similar credentials and cannot accept all of them. Should the admissions officer more often accept Black applicants than White applicants?". **Effort.** Participants will complete two self-report items assessing their effort and motivations in taking the study. The items were "Did you care about your performance in the study?" and "What level of effort did you put forth in the study?". The first item had the response options "Not at all", "Slightly", "Somewhat", "Very much", and "A great deal". The second item had the response options "No effort", "Slight effort", "Moderate effort", "Strong effort", and "Extreme effort". **Interventions** The nine successful interventions from Lai et al. (2013) will be re-tested (with no or minor revisions) in the proposed study for the longevity of their effects (for more information about the development and theoretical bases for these interventions, see Lai et al., 2013). Participants in the control condition will not complete any intervention. These nine interventions can be characterized by four categories: 1. **Exposure to counterstereotypical exemplars.** These interventions were designed to induce reductions in implicit preferences through exposure to positive Black exemplars and negative White exemplars. Five interventions from this category will be tested in the proposed study (Vivid Counterstereotypic Scenario, Practicing an IAT with Counterstereotypical Exemplars, Shifting Group Boundaries through Competition, and Shifting Group Affiliations Under Threat). 2. **Appeals to egalitarian values.** One intervention from this category, Priming Multiculturalism, will be tested in the proposed study. Participants in this intervention read a prompt advocating multiculturalism and wrote about multiculturalism’s benefits. 3. **Evaluative conditioning.** Changes in attitude can result from simple pairings of attitude objects (e.g., Black/White faces) with other valenced attitude objects (e.g., Positive/negative words). Two different paradigms will test the efficacy of conditioning in the proposed study. (Evaluative Conditioning, Evaluative Conditioning with the GNAT). 4. **Intentional strategies to overcome biases.** Performance on implicit measures can be modulated by the use of motivated strategies to override automatic bias. Two manipulations will take this approach to implicit preference reduction in the proposed study (Faking the IAT, Using Implementation Intentions).