Patricia Devine - US grants

DESCRIPTION (provided by applicant): All humans-no matter how intelligent, egalitarian, or well-intentioned-are susceptible to cognitive biases in the way they make decisions and judge others. Although these biases can operate unintentionally in opposition to one's conscious intentions, personal beliefs, and objective data, they may unwittingly perpetuate social inequalities. Unexplained disparities in R01 funding outcomes by race and gender have raised concern about bias in NIH peer review. This Transformative R01 will examine if and how implicit (i.e., unintentional) bias might occur in R01 peer review through the following three Specific Aims: Specific Aim #1. Identify the extent to which investigator characteristics influence the words and descriptors chosen by R01 peer-reviewers and how text relates to assigned scores. We will validate positive and negative grant evaluation word categories, analyze the text of a national sample of R01 reviews, and compare the grant review text for different investigator characteristics. We hypothesize that categories of words and descriptors will differ in ways that suggest implicitly different evaluation standards by applicant race and gender, even when application scores and funding outcomes are similar. Specific Aim #2. Determine whether investigator race, gender, or institution causally influences the review of identical proposals. We will conduct a randomized, controlled study in which we manipulate characteristics of a grant principal investigator (PI) to assess their influence on grant review outcomes. We will request donations of actual funded R01s and, within each grant, manipulate the PI's gender, race, or home institution. We will then invite reviewers in the appropriate discipline to review the proposals, and we will analyze written reviews and scores. We hypothesize that investigator variables will significantly influence scores and review text such that grants attributed to higher status groups (male, White, prestigious institution) will obtain better scores and text will suppor implicitly different standards of excellence. Specific Aim #3. Examine how interactional patterns among study section members promote receptivity and resistance to discussion topics and associated grant applicants. In audio- and videotapes of constructed study sections, we will investigate the real-time social interactional processes in the discussions of R01 proposals. We will employ conversation analysis to examine the delivery of initial rankings and their rationales, topic development, and the processes through which final rankings are negotiated. This research is innovative because it examines for the first time the complexities of potential bias in NIH peer review. The potential impact is threefold; this research will 1) discover whether certain forms of cognitive bias are or are not consequential in R01 peer review, 2) determine whether quantitative text analysis is a useful measure of implicit bias, and 3) describe and label real-tim grant reviewer interactional patterns. Taken together, the results of our research could set the stage for transformation in peer review throughout NIH.

DESCRIPTION (provided by applicant): All humans-no matter how intelligent, egalitarian, or well-intentioned-are susceptible to cognitive biases in the way they make decisions and judge others. Although these biases can operate unintentionally in opposition to one's conscious intentions, personal beliefs, and objective data, they may unwittingly perpetuate social inequalities. Unexplained disparities in R01 funding outcomes by race and gender have raised concern about bias in NIH peer review. This Transformative R01 will examine if and how implicit (i.e., unintentional) bias might occur in R01 peer review through the following three Specific Aims: Specific Aim #1. Identify the extent to which investigator characteristics influence the words and descriptors chosen by R01 peer-reviewers and how text relates to assigned scores. We will validate positive and negative grant evaluation word categories, analyze the text of a national sample of R01 reviews, and compare the grant review text for different investigator characteristics. We hypothesize that categories of words and descriptors will differ in ways that suggest implicitly different evaluation standards by applicant race and gender, even when application scores and funding outcomes are similar. Specific Aim #2. Determine whether investigator race, gender, or institution causally influences the review of identical proposals. We will conduct a randomized, controlled study in which we manipulate characteristics of a grant principal investigator (PI) to assess their influence on grant review outcomes. We will request donations of actual funded R01s and, within each grant, manipulate the PI's gender, race, or home institution. We will then invite reviewers in the appropriate discipline to review the proposals, and we will analyze written reviews and scores. We hypothesize that investigator variables will significantly influence scores and review text such that grants attributed to higher status groups (male, White, prestigious institution) will obtain better scores and text will suppor implicitly different standards of excellence. Specific Aim #3. Examine how interactional patterns among study section members promote receptivity and resistance to discussion topics and associated grant applicants. In audio- and videotapes of constructed study sections, we will investigate the real-time social interactional processes in the discussions of R01 proposals. We will employ conversation analysis to examine the delivery of initial rankings and their rationales, topic development, and the processes through which final rankings are negotiated. This research is innovative because it examines for the first time the complexities of potential bias in NIH peer review. The potential impact is threefold; this research will 1) discover whether certain forms of cognitive bias are or are not consequential in R01 peer review, 2) determine whether quantitative text analysis is a useful measure of implicit bias, and 3) describe and label real-tim grant reviewer interactional patterns. Taken together, the results of our research could set the stage for transformation in peer review throughout NIH.

Breaking the prejudice habit: An evidence-based, campus-wide intervention effort, Automatically activated stereotypes can lead to biased thoughts, feelings, and behaviors, even among people whose values strongly oppose bias. Like unwanted bad habits, these unintentional biases can be extremely diffi- cult to control. My early pioneering work forms the foundation of our contemporary understanding of how people who consciously renounce prejudice have unintentional or implicit biases that leads them to be unwittingly com- plicit in the perpetuation of ongoing intergroup disparities. The specter of unintentional discrimination has inspired widespread calls from researchers, scholars, and public policy officials to develop and implement effective inter- ventions to reduce and eliminate the negative effects of unintentional bias. Reducing bias and increasing representation of racial minorities and women in science, technology, engineer- ing, and math (STEM) fields has been a top priority for NIH and nearly every other scientific organization. In the 2016-2020 Strategic Plan, NIH identifies the lack of scientific workforce diversity as a critical barrier to progress. Stereotypes and unintentional biases are key obstacles to scientific workforce diversity, making it harder for members of under-represented groups to pursue careers in science on multiple fronts. Unintentional biases can similarly influence those who are already in the scientific workforce, making decisions about who to fund, mentor, admit, or hire. They also influence potential scientists-to-be, who need to determine whether a career in science is right for them and whether they ?fit? in the scientific workforce. Many of the responses to the clarion call for strategies to reduce the impact of unintentional bias have taken the form of interventions that are not evidence-based. And, though well-intentioned, these efforts at best do not work and very often make bias problems worse. Effectively solving social problems, like that of unintentional bias, requires evidence-based interventions that produce changes that endure and affect real-world outcomes. The sole intervention that has been empirically demonstrated to produce lasting, meaningful bias reductions is my prejudice habit-breaking intervention. This intervention's initial success is rooted in decades of my research de- veloping the prejudice habit model, which proposes that ?breaking the prejudice habit? can be achieved through a combination of awareness, concern, and effort. In several randomized-controlled studies, my colleagues and I have demonstrated this intervention's long-term effectiveness. As a MIRA investigator, I would be able to 1) continue the basic bias research that forms the basis of this intervention, 2) expand the field-testing and refinement of the intervention, 3) assess a range of behavioral out- comes affected by the intervention, and 4) create automated, easily-exportable versions of the intervention that can be widely disseminated.