"In line with the basic tenets of theories of associative learning and representation, the IAT rests on the assumption that it ought to be easier to make the same behavioral response (a key press) to concepts that are strongly associated than to concepts that are weakly associated (Greenwald et al., 1998)." NGB2005.PSPB.pdf (washington.edu)
Here is some examples.
In this example, the items to be categorized are displayed at the top of the screen, with “female” and “home” on the left side and “male” and “career” on the right side. The participant sees the word “she” displayed in the middle and categorizes it to the left side. The categorization is done by pressing the keyboard keys associated with the left or right side (for example, 'e' for left and 'i' for right). The words displayed in the middle include “he,” “man,” “boy,” “she,” as well as “kitchen,” “office,” “garden,” “profession,” and “marriage.”
Next, the categories are switched, with “career” and “home” reversed. The participant sees the word “office” displayed in the middle and categorizes it to the left side, but many people may take longer to respond compared to the previous categorization. This test measures whether we hold implicit stereotypes that associate women with home and men with career. Importantly, even female participants may strongly associate women with home.
The IAT can also measure implicit “attitudes.” Attitudes refer to a person’s likes or dislikes, such as liking flowers or disliking insects.
In this example, the categories “flowers” and “good” are on the right side, while “insects” and “bad” are on the left side. The words displayed in the middle include “lily,” “evil,” “gentle,” “flea,” “rose,” and “heaven.”
When the categories “pleasant” and “unpleasant” are switched in the test, participants (especially those who do not particularly like insects) will likely take longer to categorize compared to the previous classification. Of course, the opposite may occur for those who love insects.
The IAT’s ability to measure implicit stereotypes and attitudes has had a significant impact on the academic community, leading to numerous related studies and publications. The IAT can be taken on Harvard University’s “Project Implicit” website (Project Implicit (harvard.edu)). To date, the IAT has been taken at least 40 million times.
The term “Implicit Bias” is more commonly used and recognized globally than “Unconscious Bias,” likely because this test is so well-known.
Understanding what this test aims to measure and how it does so reveals that it is fundamentally different from the “unconscious bias surveys” in Japan, which use question formats like “Do you think women are suited to be leaders?”
Greenwald and Banaji, the proponents of unconscious bias, often referred to their early research memories during conferences and academic gatherings. Memory is classified into implicit memory and explicit memory based on its nature. The terms implicit and explicit are used not conceptually but based on research experience. Implicit refers to indirectly measured results, while explicit refers to directly measured results. For example, when stimuli are given to amnesiac patients who cannot remember new information, the effects may manifest in their actions even if they are not consciously aware of it. Such memories are observed indirectly and are called implicit memories.
However, it is stated that implicit measurement is not necessarily unconscious measurement, and explicit measurement is not necessarily conscious measurement. Research results suggest that both direct and indirect measurements need to be assumed to combine the influences of conscious and unconscious mental processes.
Greenwald and others named the test that measures stereotypes and attitudes indirectly as the Implicit Association Test (IAT).
If we follow their naming intention, it is preferable not to list implicit bias alongside unconscious bias. It is desirable to use the term implicit bias for implict stereotypes and implicit attitudes measured by the IAT.
References:
No comments:
Post a Comment