Implicit Association Test

[Knowing]

This is soooooo cool and sooooooooo disturbing!

I was reading <Blink> last night, and the author said he took IAT of race and turned out subconsciously he moderately associateed positive imagines to white, despite the fact consciously he considered himself totally fair, AND he was half black. He felt so disturbed that he took the test several times, turned out the same.

Out of curiosity I took the test this morning. Out of all tests provided, most I "dare not" try, like gender and career, or gender and science, or disablity, or Arab Muslin, or Religion. Because IAT test your automatic association, subconsciousness is affected so heavily by media and past that I have no confidence in my own subconscious innocence.

So I take the gay-straight/good-bad test. I figure, this is one category I should do fine. After all, I have so many close gay friends. I see gays all the time. I watch Will & Grace. I was quite ignorant of the existence of homosexuality before 23 or so, have never had strong prejudice against gay all my life-- hey I grew up in China, and GAY did not exist on TV at all. All and All, I figure, even if I prefer straight moderately over gay, it is understandable because after all, I am straight myself.

Guess what, the association tested showed my automatic association showed STRONG preference to straight people!

Wah...I don't even know myself, subconsciously!

https://implicit.harvard.edu

[Jun]

Interesting. I did the gender and family/career test. The result is slightly female--family and male--career. This is close enough to my self impression of slightly female--family and male--neutral automatic thoughts.

I suspect, though, that the design of the tests, ie the order of word position and reversal, biases the results toward more "conventional" beliefs, thus skews the testees' true subconscious toward the more "traditional" attitude.

[Knowing]

Why do you think so?

[Jun]

Here's how the test run:

1st round: Female left, Male right.
2nd round: Family left, Career right.
3rd round: Female and Family left, Male and Career right.

By the end of these 3 rounds, I am used to hitting e for female and family related words, and hitting i for male and career related words. OK? Automatic and fast.

Then in the 4th round, the order is switched so that Male and Family are on the left side, and Female and Career are on the right side. I made a far more mistake in this round and was slower, mostly because (I feel) I had to reverse the "habit" I had formed in all first 3 rounds. Remember that the order "Female left" and "Male right" occurred twice before and Female right is a new thing for Round 4. Of course it takes more effort to think and not make an error.


Also contributing to the problem is that my attention was at its highest in the first round and declined over time. It was more difficult to concentrate as the test goes on.

The test is rigged to a certain extent. It is clearly not objective.

Is the gay/straight test also designed in favor of "traditional" attitude? In other words, is the most "untraditional" order presented at the end of the test ONLY ONCE and in reverse of the first 2 "training" rounds?

Take a look at the order of each round above. You are pretty much trained to use the same finger (same side) for Female and Family for 3 entire rounds, then you are forced to reverse it at the end. It's not testing for your own subconscious. It's physical conditioning.

Just because it's Harvard-produced doesn't mean it's perfectly trustworthy...

[密斯张三]

Your data suggest a slight automatic preference for Gay People compared to Straight People.

可英文的好坏词语我还是要先在脑子里轮一轮，不够innate，所以也许还不够“下意识”，反而是practice and trial的过程。

[Jun]

People in general do not give enough credit for conscious thoughts. It's all Freud's fault.

So my point is xiaoK you are fine. You are NOT prejudiced.

[tiffany]

I think the intrinsic flaw of the test is the sequence in which they presented the test. So, I took the gender-science test twice, first time I got a strong corelation of male with science, next time no relation.
the difference was the sequence of association.

I remember somebody mentioned that when a questionire was given before a test, and ask for the gender of the test taker, the girls score lower than when the questionire was handed out after the test. As far as I remember, this is one arguement against a recent IQ statistics, stating that a boy's IQ in general is higher than a girl's.

human psychology is an interesting thing.

[Knowing]

I remember somebody mentioned that when a questionire was given before a test, and ask for the gender of the test taker, the girls score lower than when the questionire was handed out after the test.

That is also in <Blink>!

[密斯张三]

and there's a similar questionire with racial profiles, right?

[dropby]

目瞪口呆. 我用的是中文进行测试, 因为我觉得我对中文反应会比较快. 我居然对同性恋有很强的自动化的偏好. 我可是一个同性恋朋友都没有. 只认识一个同性恋同事, 还是关系很远的那种.

[猫咪头]

I was quite ignorant of the existence of homosexuality before 23 or so,

This is why. The test is right. You associate "heterosexuality" = normal, ordinary boring etc.
That you have gay friends to spice up your life is further evidence that you associate "gay" with inordinary abnormality etc.

ubconsciousness does not lie, babe.


(Am still pouting that I have no GBF)

[dropby]

我对白人有很强的自动化的偏好. 我很羞愧地承认, 这个大概是对的.

[Knowing]

The words in the test are not so much normal or abnormal, but more like good or bad. "Superb" is not so much of Normal.

Actually I took the test again and got the following comments:
Your data suggest a slight automatic preference for Gay People compared to Straight People.

I did badly at the test. If anything that test prooved my subconcious decision making ability was terrible. 1/3rd of the time I picked the wrong choice...

[猫咪头]

Now I am officially confused by the test.

I try to put a wield looking couple to "gay" but got an "X" and forced to click it to "straight".
Then my reflex to put "endurance" to "good" was marked wrong too, and was forced to put it to "bad" against my will.

All in all, the F#@%ing test feels like a power struggle to train my fingers. And in the end, it gives no score but claimed I've maken too many mistakes.

How can reflexs be labeled mistakes ?

[dropby]

Maybe you should try the Chinese one?

[camellia]

Your data suggest a slight automatic preference for Straight People compared to Gay People.