One of the presentations I made at the University of San Diego Summer Autism Conference was titled Bias, Stigma and Stereotyping and their effects on Autistic People. In the talk I explained that one tool researchers use study bias is called the Implicit Association Test or IAT . I don’t have extensive knowledge about IATs but I know a little about them because when I was an undergrad at UC Davis (a couple of years ago) I had a volunteer position as a research assistant to a professor who used IATs to study bias, specifically bias against particular ethnic groups.
My job was to interact with the research participants (who were likewise undergraduates), one of my most important duties was to get a signature from each participant on a release form that explained their rights as test subjects. I also had a little memorized statement to tell them about what they they were to do and what they should expect while taking the test (on a computer). Afterward, I was supposed to answer any simple questions that they had and tell them how to contact the professor or a post-doc to answer more complicated questions.
Occasionally the students really wanted to talk about what the material used in the experiment or about how the experiment was designed, and occasionally I had time to talk about it with them. Besides directing them to direct their questions to the researchers, I told them that if they wanted to learn more about the IATs, in general, there was a popular book called Blink, by Malcolm Gladwell, that would give them a little more information. I also told some of them about the Harvard University website called, Project Implicit which is dedicated to sharing information about the Implicit Association Test and is also used to gather data from the (anonymous) people who take the tests online (for free).
The experiment with which I was assisting wasn’t designed didn’t give the “results” of the tests to each individual participant, their data were not “attached” to them as a person with a name, but it was aggregated, more or less, with dozens of other people’s data. But, if you take an IAT on the Project Implicit site, the computer tells you your “results”. Here’s a link to the Project Implicit website:
https://implicit.harvard.edu/implicit/demo/takeatest.html
Near the beginning of this presentation in San Diego I read the following quote:
Every man has reminiscences which he would not tell to everyone but only to his friends. He has other matters which he would not reveal even to his friends but only to himself, and that in secret. But there are other things that a man is afraid to tell, even to himself, and every decent man has a number of such things stored away in his mind. – Fyodor Dostoyevsky
Dostoyevsky was writing about what people know at different levels of consciousness. I would say that he was pointing out that there are things that people believe that they would prefer not to believe, and those things are sometimes subconscious. I got the quote from the Project Implicit website.
If you read the book Blink, you may remember that he explained that everyone has biases and we don’t all like or agree with our biases. We may even have subconscious biases against people of our own group or ethnicity. Those biases might be there because of negative portrayals of people of our own group in the media and in the dominant culture we are surrounded with.
This picture is a sample of a page from one of the IATs on the Project Implicit website. If you saw this particular page
during a test you would have been instructed to place the face in the category on the left or the one on the right, and as quickly as possible. Basically, you hit the “e” key on your keyboard if you think the picture or word in the middle if the screen belongs in a category that is listed on the left of the screen, and you hit the “i” key if you think that picture belongs in a category that is found on the right side of the screen.
For this particular IAT photographs of faces of White Americans or Native Americans (all the photos of faces were from late 1800s as far as I could tell). So for this particular page you would probably want to hit the “i” key because this fellow looks like a Native American and that category is on the right side of the page at this point in the test. The photos of faces were interspersed with modern color pictures of places that were located either in America or in a Foreign country. I don’t remember which places they showed, but for instance, a picture of the Eiffel Tower, or the Great Wall of China, might be shown and if you were taking this IAT you’d decide quickly if it was a photo of an “American” place or if it was a “Foreign” place.
Project Implicit offers free access to some journal articles about IATs, as far as I could tell, they don’t have any that touch on bias and autism. But they did have a paper there called: Implicit and Explicit Stigma of Mental Illness: Links to Clinical Care. It’s by Peris et al. It is “in press.” Obviously, keeping in mind that this is not a paper about stigma and autism, but about stigma and mental illness, I think it’s still worthwhile looking at what these researchers found.
In total, 1,539 participants were interviewed online and took an IAT (not unlike what I described above, using a keyboard to categorize words as quickly as possible as belonging to a category shown on the left or right side of the screen. More than half of the participants had studied psychology. This group ranged from undergrads studying psychology up to practicing psychologists. The bulk of the participants were professionals who not only had training in psychology, but presumably lots of experience of being around people with varying sorts of mental illness. They also had a control group of people without psychology training.
If you download the paper you can get the details, but to give you a very brief idea of what they did. The experiment the participants were asked to report what they thought about mentally ill people. Quoting the paper:
Participants rated their attitudes toward “mentally ill people” on a 7- point semantic differential scale (1 = bad to 7 = good). Analogous scales were completed for two other common stereotypes about persons with mental illness: blameworthy/innocent and helpless/competent. Note, the same ratings were also made regarding “people on welfare” to match the relative comparison category used on the IAT measure. In this way, we could more readily evaluate the relationship between the implicit and explicit bias measures.
They took an Implicit Associations Test designed to look for implicit associations or biases that the participants might have that would tend to link mental illness with negative words.The test compared judgments about people with mental illness with judgments about people who are on welfare.
In the first sorting condition, items representing the category Mentally Ill People (e.g., diagnosis, disorder) were categorized with the same response key as items representing the category Good (e.g., wonderful, joyful), while items representing Welfare Recipients (e.g., unemployed, poor) were categorized with the same response key as items representing Bad (e.g., terrible, awful). In the second condition, Mentally Ill People and Bad items were categorized with one response key, and Welfare Recipients and Good items were categorized with the other.
You may have to read the whole paper and perhaps some other background information on IATs to understand how they are designed and scored in order to get the rationale behind the study design.
The participants were also shown clinical vignettes which were “DSM IV based” clinical descriptions of people with problems who could be diagnosed as mental illness. They also looked at one vignette of a person with “general psychological difficulties” that didn’t really fit into any mental illness category.
This study reported the interactions between explicit and implicit biases and “clinical decision making”. Explicit biases are ones that the participants self-reported. Implicit biases were those not stated openly, but were “implicit associations” that were revealed by the IAT.
The were some findings that I thought were possibly applicable to the experience of autistic people being treated or diagnosed in “clinical settings.” The more training a person had in psychology, and presumably the more exposure that person had to people with mental illness, the more positive view they tended to have of mentally ill people. Which is good news. The bad news, which is old news to most people, is that the controls which are meant to represent the average person on the street without particular training in psychology (when compared to those with training) had more explicit and implicit bias against the idea of mentally ill people. Mental illness carries with it a heavy load of stigma. It may seem natural enough for us to fear and stigmatize mental illness, but it’s not so fun for those with mental illness to deal with being treated badly and thought of badly simply because of a label, not necessarily because of anything in particular that they have done.
Back to the study’s findings, even though people with training in psychology tended to have relatively less bias the mentally ill, there was still enough bias among them to measure and make correlations with how they tended to see those in the clinical vignettes. Peris et al found that if a clinician openly expressed a bias against mentally ill people, that he or she would tend to give a worse prognosis for a person whose description he had read. A bad prognosis might look something like, “this person is unlikely to be able to remain employed,” or “this person is likely to harm himself or others.”
On the other hand, if the person openly expressed no bias against mentally ill people, but nonetheless showed a perhaps hidden bias against the mentally ill via the IAT, that person tended to give a better prognosis, however tended to “over pathologize” or add on extra labels, the paper called this, “over-diagnosis”. In over-diagosis the person described in the clinical vignette got additional labels that the writers of the vignettes didn’t intend. For instance, the person in a clinical vignette who had symptoms best fitting a single diagnosis of alcohol dependence might get additional inappropriate labels like, anxiety disorder or major depression.
To quote the paper’s authors: “… the finding of a link between bias and clinical decision-making is striking and suggests that negative views toward mental illness may influence clinical care, even among individuals with considerable mental health training.” Of course, the authors say that the study should be extended to see if the way the psychologists reacted to the clinical vignettes is the same as they would diagnose and give prognoses in their practices.
Coming back to autism, maybe we need the think about whether or not bias against autistic people could likewise influence clinical care of autistic people. It’s well known that parents will work to get clinicians to either give or not give particular autism spectrum diagnoses. A parent of a child with bipolar disorder may prefer that their child be diagnosed with Asperger’s since Asperger’s is still a recognized disorder or disability, but maybe it has a cachet of “genius” that bipolar doesn’t have. And if you are very much involved with autism at all you probably know that people will try to get a kid who generally would be seen as “Asperger’s” to get an “autistic disorder” diagnosis so that the kid can get more services paid for by some entity. And some parents of obviously autism spectrum kids pressure diagnosticians to take the label off their kid because they don’t want the child to be ….uhm… stigmatized. And, unfortunately some parents want to have their efforts at curing their child to be officially recognized as successful, so maybe they’d like to say, “See I gave my kid 5,000 methyl B12 shots and she had 40 rounds of IV chelation and 300 dives in an HBOT tank and it all paid off!” or maybe they’d like to say, “Selling our home and taking on extra jobs and begging money from all our relatives to pay for ABA therapy has left us in tatters financially, but we got our real boy back.” Of course, it is appropriate to re-diagnose autistic kids sometimes. Some do start out looking very autistic and end up looking close to normal, at least some of the time. You can watch a couple of talks by Deborah Fein on the MIND Institute website about this very thing. She points out that kids sometimes move from looking autistic to looking ADHD, perhaps plus OCD, or plus ODD or anxiety… which makes me wonder about how much the clinician’s biases are influencing the desire to rid the kid of a single diagnosis or PDD,nos and replacing it with mulitples like “ADHD plus social anxiety plus depression.” Hmmm?
We know that clinicians sometimes give negative prognoses for very small children, such as, “this two year old child is autistic. He will never marry, never have a job, you might as well lock him up now in an institution and get on with your lives….” and we also know that some autistics have lots of different diagnoses on top of an ASD. We might ponder about why that is, is it because of bias against autistic people and autism in general? We don’t know because, as far as I can tell no one is studying it. Autism clinicians are pretty much all assumed to be blank slates who don’t bring to the table their own biases that influence their ability to judge a situation fairly or accurately. Autistic people are, in my opinion, inherently “irritating” to most folks. We aren’t so ingratiating or charming, we aren’t precisely acting on typical social rules, generally speaking, to say the least. Still if you learn more accurate things about autism as opposed to stigmatizing misinformation such as is commonly found on the Autism Speaks website or from the mercury parent organizations where you may read that autistics destroy families and are walking toxic-waste carrying “train-wrecks,” or you may hear that autistic teens are tsunamis of fiscal devastation, thank you Stephen Shore. If you learn the true things about autism and autistic people you may actually start to like autistic people. You might think we are pretty fabulous. I think we are pretty fabulous. If you listen to people like me say things like, “autistics are pretty fabulous,” you can actually start to think differently about autism and perhaps change your biases.
I also found three other papers that I thought were interesting, they were referred to in this book chapter on IATs. (Lane, K. A., Banaji, M. R., Nosek, B. A., & Greenwald, A. G. (2007). Understanding and using the Implicit Association Test: IV: Procedures and validity. In B. Wittenbrink & N. Schwarz (Eds.), Implicit measures of attitudes: Procedures and controversies (pp. 59-102). New York: Guilford Press.) You can request this chapter from the project implicit website. It’s free.
Could similar problems caused by bias against autistic people be affecting autistic people now?
In one study, physicians who held stronger negative stereotypes of African American people were more likely to give them less adequate care after a heart attack. Or you could say, “Belonging to a group your doctor has a bias against can be bad for your health.”
People with a bias against African Americans score lower on a test of intelligence in the presence of an African-American person. Or, if you are in the presence of someone who has a bias against you you can expect them to behave less intelligently.
Stronger implicit stereotyping of members of a group was associated with more negative judgments of ambiguous actions by a member of that ethnic group. Or, if people have a bias against you they will tend to interpret your neutral actions as negative.
What biases or attitudes about the autism spectrum, autistic children and autistic adults might we be hiding from ourselves? Is this important to ask?
Left Brain Right Brain 
Recent Comments