## Method ## *Inclusion Criteria* Valid meta-analysis requires careful consideration of which studies are relevant to the research question and which studies are not. Given our focus on procedures that causally change implicit bias, we set the following inclusion criteria: (1) The study is a between-subjects experiment. This led to the exclusion of studies that used correlational or quasi-experimental designs (e.g., Rudman, Ashmore, & Gary, 2001) and within-subjects manipulations (e.g., Wheeler & Fiske, 2005). (2) The study includes an implicit measure that is administered after the onset of the experimental manipulation. Implicit measures were defined as measures of associations between concepts that do not require the participant to actively bring to mind the target association. Measures for which the manipulation began during task instructions or practice trials (e.g., Foroni & Mayr, 2005) and for which the manipulation extended into the measure (e.g., Huntsinger et al., 2010) were considered eligible. (3) The implicit measure assesses a pre-existing association. We defined a “pre-existing association” as an association that either should theoretically be present, or has been empirically demonstrated as present, within a target population prior to the onset of the study. (4) The experimental procedure must fit into a single procedure category. Procedure categories were created iteratively with the goal of capturing the breadth of approaches in the literature. This iterative process meant that the included procedure categories (and studies) changed during the coding process. Procedures that fit into multiple categories or did not fit into any categories were excluded. If a study only had one condition remaining after exclusions, the full study was excluded. For more information about this criterion, see the section labeled “Experimental procedures” below. (5) The study is reported in English. This criterion led to the exclusion of studies that were not written in English. ---------- *Literature Search* Studies were obtained through several methods: (a) PsycINFO and Web of Science seraches, (b) direct requests for relevant studies through email and the SPSP listserv, and (c) an additional 115 articles from an unpublished meta-analysis of the malleability of implicit intergroup bias. *Database Search*. We searched the literature at the start of the project (September 2012) and near the close of active data collection (June 2014). We used a combination of search terms reflecting our various inclusion criteria. A study needed to be be fulfill all search criteria to be included in our initial sample of studies for the meta-analysis. **1.) To locate research using implicit measures:** "Implicit attitude*" OR "implicit stereotype*" OR "implicit self-esteem*" OR "implicit self-concept*" OR "Implicit evaluation*" OR "implicit self-evaluation*" OR "implicit association*" OR "implicit identit*" OR "implicit prejudice*" OR "Automatic attitude*" OR "automatic stereotype*" OR "automatic self-esteem*" OR "automatic self-concept*" OR "automatic prejudice*" OR "Automatic evaluation*" OR "automatic self-evaluation*" OR "automatic association*" OR "automatic identity*" OR "implicit preference*" OR "automatic preference*" OR "implicit belief*" OR "implicit bias*" OR Any Field: "Implicit Association Test" OR "Implicit Attitude Test" OR "Evaluative Priming" OR "Go/No-Go Association Task" OR "Affect Misattribution Procedure" OR "Lexical Decision Task" OR "Semantic Priming" OR "Single-Target Implicit Association Test" OR "Approach-Avoid Task" OR "Breadth-based Adjective Rating Task" OR "Evaluative Movement Assessment" OR "Extrinsic Affective Simon Test" OR " Implicit Association Procedure" OR " Implicit Relational Assessment Procedure" OR "Name-Letter Effect" OR "Recoding Free IAT" OR "Sequential Priming" OR "Single Block IAT" OR "Sorting Paired Features" OR "Stimulus Response Compatibility Task" OR "Brief Implicit Association Test" OR "IAT" OR "Single-Category Implicit Association Test" OR "affect misattribution procedure*" OR "approach avoid* task*" OR "approach avoid* test*" OR "breadth based adjective rating" OR "affective priming" OR "bona fide pipeline*" OR "bonafide pipeline*" OR "extrinsic affective simon" OR "go no go association*" OR "implicit association* procedure*" OR "implicit association* test*" OR "implicit association* task*" OR "implicit attitude* test*" OR "implicit attitude* task*" OR "lexical decision task*" OR "lexical decision making task*" OR "lexical decision test*" OR "lexical decision making test*" OR "implicit relational assessment procedure*" OR "name letter effect*" OR "name letter task*" OR "name letter test*" OR "sorted paired features task*" OR "sorted paired features test*" OR "stimulus response compatibility task*" OR "stimulus response compatibility test*" OR "linguistic intergroup bias*" OR "stereotypic explanatory bias*" OR "shooter task*" OR "shooter test*" OR "shooting task*" OR "shooting test*" OR "simple association* test*" OR "single association* test*" OR "simple association* task*" OR "single association* task*" OR "weapon* identification" OR BIAT OR IAT OR GNAT OR IRAP OR LDT OR RF-IAT OR ST-IAT OR SC-IAT OR SRCT OR SB-IAT OR BIATS OR IATS OR GNATS OR IRAPS OR LDTS OR RF-IATS OR ST-IATS OR SRCTS OR SB-IATS **2.) To locate studies focusing on associative change:** malleab* OR chang* OR influenc* OR moderat* OR reduc* OR increas* OR shift* OR alter* **3.) We also included additional search criteria for language (i.e., English-only) and studies published between 1995-2015.**