As I said I would last week, I have dived into reading the research of others on implicit bias. I wanted to use this post to share some of what I learned. The study I read is called Community Perception: The Ability to Assess the Safety of Unfamiliar Neighborhoods and Respond Adaptively. Unlike many of my other sources, this one was focused on the implicit biases people create from neighborhoods and the buildings within them. The study had two goals, to see if the subjects could accurately estimate the safety of an unknown neighborhood, and to see if they were using indicators of disorder to do so.
The study is based on an interesting theory called disorder theory or ‘broken windows’ theory. It is based on the premise that neighborhoods readily display their safety through physical indicators, such as broken windows. This is done because social or physical disorder within the community will lead to a lack of maintenance of public structures. The participants of the study were then shown photos of different neighborhoods followed by questions.
The method used that is most applicable to me was the Sequential Prisoner’s Dilemma. This game has similar payoffs to a normal prisoner’s dilemma game, except that the decisions were made at different times. Meaning that one player would go first, and the second player would make their decision with full knowledge of what the first person had chosen. Two conclusions were drawn from this. First, the likelihood of the first player to give an offer of cooperation was correlated with the social quality of the neighborhood, which is what the researchers expected. What was surprising is that when given an offer of cooperation there was still a correlation between cooperation and social quality, despite the fact that the trustworthiness of the other player now no longer mattered.
The other method used was something you probably all know, but have never heard called by it’s actual name, a Likert Scale. It is a five-point scale used in a response to a statement that has five possible responses, strongly agree, agree, neutral, disagree, and strongly disagree. The Likert Scale is widely used in a variety of surveys and studies due to its simplicity.
Moving on from the methodology, I also noted my complete lack of knowledge in modeling, something I am going to need to work on in the future. This study used something called Hierarchical Linear Modeling, which they used to show a correlation between the social qualities and different participant responses. Furthering my understanding of modeling is going to be imperative going forward since without it I won’t be able to do anything with all of my data.
Thanks for reading.
For those curious to learn more about the study and its findings: