https://doi.org/10.5061/dryad.cz8w9gj8t

There are a total of 5 datasets available (Studies 1a, 1b, 3, 4a, 4b) each collected by the researchers from online survey platforms. All data files are .sav files. We recommed using SPSS or RStudio to work with the data. We provide our code using RStudio and a codebook with the name of all variables in each dataset.

Description of the data and file structure

Study 1a and Study 1b utilized a within-subjects experimental design (S1a: N=191; S1b, preregistered: N=337, 50% White participants, 50% Black participants) where samples of U.S. citizens recruited from MTurk.com and Prolific Academic read nine examples of racial disparities, three each from the domains of health, economics, and belonging. After each example, participants reported whether the disparity was unjust and fair (reverse-coded; 2-items averaged to create a perceived injustice scale). Participants also indicated their agreement (1=strongly disagree, 7=strongly agree) that they “personally want to take action to reduce the given disparity” (general action). In S1b, participants additionally indicated how likely (1=extremely unlikely, 7=extremely likely) they were to engage in specific actions: donating, engaging in a protest, and sharing information on social media to reduce the disparity (concrete actions, 3-items averaged to create a concrete action scale). To assess the effects of domain (health, economics, belonging; S1a and S1b) and participant race (S1b only) on action support, we conducted multi-level models with domain condition (S1a and S1b) and participant race (S1b) as fixed effects and participant and scenario type as random effects (intercept-only). To assess the mediating factor of perceived injustice we used within-subject mediation (JSmediation package).

Study 2 (preregistered) used Facebook’s ad platform and thus the data and analyses are collected and run by Facebook (the researchers do not have access to the data). Screenshots of the findings provided by Facebook are provided in OSF: https://osf.io/5bps2/?view\\\\_only=cdb88ebca7bf41a4a25d6a9fea9926df.

Study 3 (preregistered) is from a sample collected from NORC’s AmeriSpeak Panel via the Time-sharing Experiments for the Social Sciences Short Studies program of 1,550 U.S. residents. Participants viewed a short infographic highlighting either health-, economic-based, or belonging-based racial disparities (between-subjects design). Then all participants reported on their agreement (1=strongly disagree, 7=strongly agree) of how unjust (“These disparities are unjust”) and morally sacred (2-items averaged to create a scale; e.g., “These racial disparities involve issues or values which should never be violated”) the issue was as well as their support for general policies to reduce it (2-items averaged to create a scale, e.g., “Politicians need to prioritize creating policies that reduce these racial disparities”). To analyze the data, we conducted an ANOVA and planned contrasts to test if the type of disparity influenced policy support. To assess the proposed serial mediation (highlighting health (vs. economic) disparities elicited support for mitigating policies due to perceptions that the issue is morally sacred which in turn, induced perceived injustice), we used process mediation model 6 (process package).

Study 4a (N=490, 76% White) and Study 4b (preregistered, N=1,088, 74% White) had U.S. citizens recruited from Prolific Academic indicate their support (1=strongly oppose, 7=strongly support) for the same fiscal federal policy (policy of interest: increasing taxes by 0.5% for all Americans). Specifically, participants indicated their preferences for a total of eight policies in S4a (exploratory) and two policies in S4b. Crucially, participants were randomly assigned to one of two framing conditions in which they either read that these policies would help to reduce health-based or economic-based racial disparities. To analyze the data, we conducted independent samples t-tests to test if disparity framing influenced policy support.

Sharing/Access information

This is a section for linking to other ways to access the data, and for linking to sources the data is derived from, if any.

Links to other publicly accessible locations of the data:

• Data and code are also provided on OSF (https://osf.io/5bps2/?view\\\\_only=cdb88ebca7bf41a4a25d6a9fea9926df)

Data was derived from the following sources:

• S1a was collected via MTurk.com; S2b was collected via Prolific Academic; S3 was collected via NORC’s AmeriSpeak Panel via the Time-sharing Experiments for the Social Sciences Short Studies; S4a and S4b were collected via Prolific Academic.

Code/Software

R code to analyze all studies is available and uploaded here (titled: HealthDisparities_Syntax.R) and via OSF: https://osf.io/5bps2/?view\\\\_only=cdb88ebca7bf41a4a25d6a9fea9926df