Unconscious Bias and Project Implicit

This week I will attempt to find sources that discuss the premise of my research topic and, if possible, some research that supports my hypothesis.

My premise: There is bias in the classroom, be it conscious or unconscious, related to demographic factors such as race, ethnicity, age, and gender.

My hypothesis: Using avatars for students and teachers to represent themselves in an online classroom environment can, at least in part, offset some of the bias.

I did some research into general bias in the classroom and found a wonderful study done by the California Teachers Association Institute for Teaching. The conclusions were as follows:

Social psychologists and social scientists have found that all of us, regardless of who we are, have cognitive biases that influence how we perceive and make decisions about other people. The behavior of human beings is often guided by racial and other stereotypes of which we are completely unaware. Simply put, those things that all of us are subject to (e.g. stereotypes, prejudices) that we are unaware of which lie deep within our subconscious. This unconscious stereotyping and prejudice reveal our "implicit attitudes" of various subjects including race and even the ability to learn. Research has shown that we develop and use cognitive shortcuts and generalizations to make sense of an overload of incoming information in order to navigate the world around us. Racial stereotyping is one method that is frequently relied upon to make decisions about people and the world. http://www.teacherdrivenchange.org/teacherdrivenchange/2008/08/unconscious-bia.html)

The study used something called the Implicit Association Test (IAT) which presents a method that demonstrates the conscious-unconscious divergences “much more convincingly than has been possible with previous methods.” The test was created by Project Implicit which is a Virtual Laboratory for the social and behavioral sciences designed to facilitate the research of implicit social cognition. These are the feelings and evaluations that are not necessarily available to conscious awareness, conscious control, conscious intention, or self-reflection.

Project Implicit comprises a network of laboratories, technicians, and research scientists at Harvard University, the University of Washington, and the University of Virginia. The project was initially launched as a demonstration website in 1998 at Yale University, and began to function fully as a research enterprise following a grant from the National Institute of Mental Health in 2003. (http://projectimplicit.net/about.php)

This new method is called the Implicit Association Test, or IAT for short. There were 15 Demonstration Tests offered and I chose the following tests: Gender-Science, Race, and Age before I got bored. The tests did not seem significantly unbiased (no pun intended) enough for me to put 100% faith into the results. However, Project Implicit is a respected Virtual Laboratory with world-wide connections, my personal opinions notwithstanding.

For each test, the test subject was given a stream of images or words. For example, for the Gender-Science Test, the first stream was a list of words that signified male or female (Mother, Father, Aunt, Uncle, Son, Daughter, etc.). Male might be the heading on the left side of the screen and Female on the right side. The ‘e’ key would signify the left side of the screen and the ‘i’ key signified the right side. If the word was a male word, the subject would press the ‘e’ key; if female, the subject pressed the ‘i’ key. In the next set of images, Science was one category and Liberal Arts was the other. The Science category would be listed under Male on the left side and Liberal Arts under Female on the right side. Again, as the subject was presented with words (such as Physics, History, Art, Geology), the ‘e’ key was pressed to signify a Science and the ‘i’ key to signify a Liberal Arts category. The next two or three screens presented the same images or words but the placement (left side or right side) would be changed. Male might be paired with Liberal Arts and Female with Science. It appears that the researchers believed that the number of wrong keypresses by the test subject and what those wrong keypresses were would indicate a level of unconscious bias.

Results for all the tests I took were given at the end and, if I can generalize from the ones I took, it appears that more than 50% of the population who took these tests had the bias which is commonly expected: the assumption that men are more prone and more able to do science/math than women, that blacks are more often associated with negative words, that people tend to prefer youth over old age.

While I don’t doubt the results, I do doubt that these tests proved these things. As a test subject myself, I “learned” where to put each category from the first round of images. It was difficult to unlearn the tendency to put category X on the left side the next time around. And since the first round of images consistently paired one thing (Male or Black) with its “expected” preference (Science or negative words), I learned to place science on the left and had to unlearn it the next time around when science was on the right side, under Female. I’m not sure I am explaining all these clearly so, if you’re interested, take one of the tests: https://implicit.harvard.edu/implicit/demo/selectatest.html

However, Project Implicit is an impressive project, with very interesting objectives, and has Web sites around the world, The more I read about it, the more I think I need to do more research into it. Next blog!

References:

Unconscious Bias in the Classroom by California Teachers Association Institute for Teaching, August 15, 2008. At: http://www.teacherdrivenchange.org/teacherdrivenchange/2008/08/unconscious-bia.html (retrieved April 15, 2009)

Project Implicit at http://projectimplicit.net/about.php (retrieved April 15, 2009)