Using Social Signals to reduce Pluralistic Ignorance towards Sexual Violence on College Campuses

This study is considered a nudge intervention. A “nudge” is defined as an intervention that changes the choice environment in which people find themselves without altering their choice set (the array of options that they may feasibly choose) in order to encourage people to behave in a more optimal (i.e prosocial, money saving, altruistic) way. This particular intervention will accomplish that through the distribution of green wristbands indicating those who have made personal commitments to be active bystanders in situations of potential sexual violence.

A large barrier to the prevention of sexual violence in public situations is the issue of pluralistic ignorance, that is that humans will take cues from those around them to inform what they should do. If the overriding view of those around them is to do nothing, someone will be inclined to also do nothing against their better judgement. An elegant example of this lies in Latane and Darley (1968), where confederates who acted like nothing was wrong induced similar behaviour in participants even when the threat was apparently dire, such as smoke beginning to fill a room. The introduction of personal commitment wristbands aims to change both individual and collective action to normalize active intervention. Individually, the wristband will invoke cognitive dissonance by forcing the wearer to make their private commitment to active intervention available for all to see on their wrist. If they then fail to act, their behaviour is inconsistent and they will undergo some cognitive strain as predicted in Festinger (1962). Collectively, the wide proliferation of wristbands should create some normative change whereby the campus at large is expressing not only intolerance for sexual violence but also support for victims which I predict will deter perpetrators and empower those who find themselves in potentially dangerous situations.

In addition to the wristband itself, two additional variations will be trialled in this study to better understand the influence of the wristbands and the most effective way to deploy them. The first of these variations lies in the framing of the wristbands, with either loss-framed or gain-framed messaging on the bands. Following from the research of Kahneman and Tversky (2013)’s Prospect theory, the framing of messages can influence one’s appetite for risk and, subsequently, their actions. The message of “be the difference” (gain) vs “don’t stand around” (loss) could induce different bystander reactions and as a result will be looked at on the wristbands.

The second variation that will be tested is how widespread the bands are in order to try and parse whether change is coming from individual or collective action. On some campuses, only 500 wristbands will be distributed whereas on others the wristbands will be available to the entire undergraduate population. If the effects that are observed are driven by collective action in addition to individual action, I would expect the reduction in sexual violence to be above and beyond the proportional difference in wristband proliferation. The 500-wristband case provides a baseline for the influence of some wristbands over the control while not providing enough wristbands for true normative change. The full proliferation condition, however, carries with it the potential for campus wide attitude change in addition to the call to action the wristband carries by itself.

There are two major outcome variables for this intervention. The first is the number of instances of sexual violence which will be measured in two ways, and the second is an index of norms surrounding sexual violence. The instances of sexual violence will be taken from Title IX reports received by the university over the period in question. In order to control for inconsistencies in the process of Title IX reporting between universities, the surveys that will be distributed to students at the beginning and end of the term will ask for reports of sexual violence prevalence using a list methodology and then asking for direct reports. The survey also contains a measure of norms surrounding sexual violence so the researchers can monitor how those shift over the course of the term. Changes in responses to these surveys between the beginning and end of terms, above the trend seen in the synthetic control for the study, will serve as outcome variables. While these surveys are not strictly part of the intervention itself, they will be used to determine the success of the intervention, in conjunction with Title IX reports received by the university through official monitoring channels.

2024-09-01

2025-06-01

Title IX reports and formal complaints, normative changes as measured by distributed survey responses

The first step of this study is the recruitment of participating universities and the assignment to one of the treatment arms (small vs large, gain vs loss, control). From there the researchers will be in contact with the universities in the summer of 2024 to communicate procedure, expectations, and to send materials (posters and bands) for their treatment condition.

Initial contact with universities will be made to the Title IX office and Title IX coordinator at each university, leveraging any personal connections to help establish contact. The reason for choosing the Title IX office at each school is due to the fact that a) each university is required to have a Title IX office, which standardizes the process, and b) the Title IX office holds statistics on student reports of sexual violence on campus. This will be one of the primary data sources to measure the reduction in sexual violence. Title IX reports are further subcategorized by type of violence (assault, misconduct, and harassment, among others), allowing the researchers to look at the effect of the intervention on each subcategory as well as by looking at changes in sexual violence more generally. In case there are any potential confounds such as stigma surrounding Title IX reporting, reporting form heterogeneity, etc. a secondary outcome measure will be self-reporting numbers taken from the baseline and endline surveys. In using these two measures, the researchers should be able to construct a robust measure of the outcome variable of interest, the prevalence of sexual violence on campus.

Before the university is chosen for participation, the researchers will require written permission from both the Title IX office and, where relevant and required, the central university administration. Due to the small number of treatment universities this can be handled on a case-by-case basis to allow the researchers to navigate the bureaucracy of individual institutions, but written permissions will be a requirement for participation.

This study will be split into two waves of rollout, in order to simulate a pilot and allow for the researchers to better understand the potential issues with distribution or study design before implementing it at all universities. The first stage will involve two universities in the fall semester of 2024 (August 2024-December 2024), one control university and one treatment university. The researchers will observe how the study design works for this initial rollout to inform any changes that may be necessary before distributing the wristbands to all treatment universities. The second wave of this experiment will occur in the second semester of that year (January 2025-May 2025) with remaining universities.

At the beginning of the term, the baseline survey will be sent around to the students at the university for completion, with 1 week allotted for its completion. It will be a Qualtrics form sent out to all students via university email by the Title IX office. This email will encourage students to respond for the sake of the research, but also will mention the potentially distressing nature of the questions and signpost resources for those who may feel distressed. This email will be written in conjunction with the Title IX office at each university to ensure it abides by all legal and university-imposed requirements. The survey will take approximately 10-15 minutes to complete. All questions in the survey will be optional, apart from a question that requires a university email address. This question will be used to verify that the participants attend the university in question.

At the end of this week, the volunteers (~5-10 per institution) in the economics/psychology/sexual assault prevention departments of the university will set up tables in busy thoroughfares of the university as well as place posters around the campus. To obtain a band, a student must sign an informed consent form and a “personal commitment” pledge to be an active bystander as well as a short description of what an active bystander is. At universities in the “large” condition, replacement bands will be on hand for the duration of the semester for those who lose or break theirs. A similar pledge must be signed to receive a replacement band. At the end of the semester (December 2024 for wave 1, May 2025 for wave 2) the endline survey will be distributed by the same means as the baseline. Both surveys are attached to the end of this document. The questions are similar for normative questions but the endline survey does contain some specific questions about uptake and usage of the bands. At this point, the data obtained through the surveys will be used in addition to secondary data obtained by participating universities of official sexual violence reports to complete data analysis. The secondary data required will just include the date of a completed report and the number of reports received, with no identifying information or details attached. All data will be anonymized before it is used. Both surveys will take approximately 10-15 minutes to complete. The surveys will require a university email to be completed but will not store this information outside of verification purposes.

Over the course of the trial, the researchers will signpost resources for complaints or comments about the research via posters and the initial survey email sent to students. The researchers will monitor these complaints or comments with the Title IX office to better understand what behaviours the students are experiencing through the trial and to discuss about the continuation of the trial. The Title IX office and the university administration will have the power to end the trial early if they determine that the risks and behaviours observed are creating sufficient negative externalities to end the trial.

Further monitoring will continue at universities through their regular channels for dealing with sexual violence. Each of these departments that could receive information regarding the progress of the trial and any negative side effects will be made aware of the reporting mechanisms available to the researchers and the Title IX office as part of this trial.

At control universities, the surveys will be distributed in the same way but no bands will be passed out.

The data analysis will follow a synthetic control approach. The control universities will be weighted to provide a single control comparison point with matched characteristics with the trial universities. They will be compared on normative changes over the course of the semester as well as the change in sexual violence reporting using official channels.

Not available

Randomization will be done by a computer in the office

University Level

No

20-40 Universities

Approximately 80,000 students overall. The exact number will be dependent on which schools participate in the study exactly, which is yet to be determined.

8 treatment schools (2x2 factorial design of gain vs loss message, wide vs narrow distribution, 2 schools in each arm), 12-32 control universities that will create a synthetic control for the existing units.