Early Research Prototype
The first step to help men become more empathetic towards women is to make them aware of their gender bias. I did some research to find tests or games designed to uncover implicit bias. I found one example that forces a person to make a snap judgement between to images of the topic being tested and the other example is an ongoing study by Harvard called Project Implicit. Their goal is “to educate the public about hidden biases and to provide a “virtual laboratory” for collecting data on the Internet.” The way their test worked was by determining how quickly you could place an “item” into its correct “category” when its item was flashed on the screen. So, if the four categories are Male, Female, Career and Family, each would have corresponding “items” such as Bob for male, Janet for female, Salary for career, and Children for family. Categories would be placed on each side of the screen either singularly or as a pair, and you would use the “e” and “i” keys to place the item into the correct category. I took inspiration from both of these methods to develop my own research prototype to reveal implicit bias.
My design would use a measure of time (1 minute), but as a mechanism to encourage snap judgment, and instead of choosing between only two images, the respondent would choose between multiple images and make a choice of three. I wanted to choose professions that people may revere and do not necessarily know what the qualifications for the position in order to make the choice solely about who the person is in the image. My assumption was that the choices made would be overwhelmingly 2 to 1, man to woman regardless of the chosen profession.
My first test used astronauts as the profession with the simple question of “Who gets to go to Mars? Choose three.” However, the image you see above is the 2nd iteration. The first test had 12 astronauts (6 male, 6 female) to choose with similar colored backgrounds and an American flag on the left side while a small scale replica of a space shuttle and rocket was on the right side. They were also of mixed racial backgrounds of White, Black and Asian. The first problem was that including different races would probably skew my data since I only wanted to test for gender bias and not racial bias. The second problem was that one of my first participants made his choices based on the background color (blue) of three astronauts, and that had nothing to do with gender. So I made the second and current prototype you see above which uses no background color, removal of superfluous objects and uses only one race because most astronauts (unfortunately) have been white and continue to be, and also I needed a large amount of astronauts and they tend to be white.
My preliminary results from this early research prototype were really intriguing. I did not just test it on men but also women, and the overwhelming majority (including all the women, about 5) chose 2 male astronauts and one female astronaut. The two instances that did not happen were by men, but I saw them as an anomaly because one did not take the time to look at all of the images and the second male participant knew exactly what I was testing for and possibly made his choices on what he thought I was looking for as the right choice.
My second research prototype for gender bias is designed to also test for trustworthiness in gender. I chose to use news anchors. However, after testing this once, I realized that using well known anchors was the wrong way to go because it leaves to much room for pre-existing bias based on how well they recognize some of the options. So, I am going to iterate and use anchors from small local broadcasts to minimize the chance that they will be recognizable.
I still want to make one or two more prototypes to test but haven’t yet decided what they will be.
