Gaming for Good: Evaluating an Interactive Online Bystander Intervention

The bystander intervention program is provided online on a proprietary website. It consists of (1) an interactive film and (2) a subsequent quiz. Both the film and the quiz follow the gamification approach by using game principles and design elements. First, the interactive film puts the user in the position of a witness who observes a potentially violent situation. After each sequence of the film, a pop-up window asks the user to choose one of two options for action. Depending on these choices, the plot of the film takes a different course. At the end of the film, the user receives feedback containing an assessment of his decisions, recommendations for optimal behavior in a similar situation, and a score.

Second, the subsequent quiz consists of six questions, each relating to the content of the film. For example, the user is asked for offender characteristics and progression of events. After each question, the user receives immediate feedback containing an assessment of his answer, the correct answer, and a score. At the end of the quiz, the viewer receives a final feedback containing an assessment of his choices and a total score.

2020-05-12

2020-07-31

1) Willingness to intervene;
2) Actual donation to victim support

1) As with most bystander intervention programs, the intervention in analysis aims to increase the participants’ willingness to intervene in situations with a high risk of violence to occur. We thus use the willingness to intervene as our first primary outcome. We measure the willingness to intervene using a set of items each reflecting a specific bystander helping behavior. Respondents are asked for their likelihood to engage in these helping behaviors on a 7-point scale.

2) The final aim of the intervention is to instigate actual behavior. However, due to ethical and organizational constraints, we cannot observe actual bystander helping behavior. As our second primary outcome, we thus use an incentivized measure to approximate actual behavior. This measure is the actual donation to victim support. Respondents receive a compensation for participating in our experiment. At the end, they are asked, which share of their compensation they are willing to donate on a scale from zero to 100%, with 10% increments.

The average effect allows for a general assessment of the policy. To consider potential effect heterogeneity, we consider age, gender, family and socio-economic status as well as living in urban versus rural areas as subgroup categories that can speak to the policy goal of violence prevention.

1) Bystander efficacy;
2) Barriers to intervention (Detection, interpretation, assumption of responsibility, perception of skills, acceptance of negative social consequences, empathy, indignation, anticipated guilt);
3) Parameters of the Reasoned Action Approach (Behavioral intention, attitudes, perceived norms, perceived control)

To analyze the mechanisms underlying the effect of the intervention in analysis on the willingness to intervene and the incentivized behavior, we use a set of secondary outcomes:

1) Following most bystander intervention programs, the intervention aims to increase the participants’ willingness to intervene by teaching the skills for safe and effective intervention. Learning these skills should increase participants’ self-efficacy. Thus, we use bystander efficacy as our first secondary outcome. Therefore, we employ the same set of items as in our first scale and ask respondents’ for their self-confidence in performing these behaviors on an 11-point scale.

2) According to the bystander intervention model (Latané & Darley, 1968, 1970) a bystander has to overcome five barriers before she or he will engage in helping behavior: (1) Detection of the emergency situation, (2) interpretation of the situation as an emergency, (3) assumption of own responsibility, (4) perception of having the necessary skills, and (5) decision to intervene. Our intervention teaches the skills necessary for detecting and interpreting situations as emergencies. Furthermore, it fosters participants’ assumption of responsibility. After the intervention, participants should thus perceive to have the skills necessary for intervention. Finally, when making the decision whether to help, bystanders conduct a cost-benefit analysis weighing the costs and benefits of intervening against the costs and benefits of non-intervening. A higher acceptance of negative social consequences should decrease the costs of intervening. In reverse, empathy, indignation and anticipated guilt should increase the costs of non-intervening. Each of these factors should lead to a positive decision to intervene. We measure each of these parameters using scales with one or two items each

3) Following the reasoned action approach (Fishbein & Ajzen, 2011), the intention to perform any given social behavior is the single best predictor of an individual actually performing that behavior. Intention in turn is a function of the attitudes towards performing the behavior, perceived social norms with respect to the behavior, and perceived control over performing the behavior. We measure behavioral intention, attitude, perceived norms, and perceived control with respect to a specific bystander helping behavior using one or two items each.

We conduct a randomized controlled trial (RCT) randomly assigning the participants into two treatment groups and two control groups. The treatment groups participate in the interactive film and the subsequent quiz. The control groups receive no treatment. The two treatment groups (and the two control groups) respectively differ in the way that the first primary outcome (i.e. willingness to intervene) is measured. Thereby we want to control for priming effects.

Randomization done in office by a computer.

Individual

No

1,600 individuals

1,600 individuals

800 indviduals control, 800 individuals interactive film and quiz