There’s a clip floating around Tik Tok from Devon Price’s Unmasking Autism.
Sasson and colleagues (2017), for example, found that neurotypical people quickly and subconsciously identify that a stranger is Autistic, often within milliseconds of meeting them. They don’t realize that they’ve identified the person as Autistic, though; they just think the person is weird. Participants in the study were less interested in engaging in conversation with Autistic people and liked them less than non-Autistics, all based on the brief moment of social data. It’s also important to point out that the Autistic people in the study didn’t do anything “wrong”; their behavior was perfectly socially appropriate, as was the content of their speech. Thought they tried their damnedest to present as neurotypical, their performance had some key tells, and was just slightly “off,” and they were disliked because of it.Unmasking Autism, pg 185
It’s a powerful clip. I remember hearing Dr. Price’s words echo as I walked away from another social interaction feeling awkward and defeated. This clip out of context really negatively affected me. And it wasn’t long before I saw others on Tik Tok post about how their mental health suffered hearing that.
For many, it seemed like a no-win scenario. If allistic people viewed my behavior as automatically less comfortable to be around, what hope did I have when socializing? My social interactions suffered under this self-fulfilling prophecy. I saw myself as unable to interact with others, which made my interactions worse.
Dr. Price’s suggestion to this was the follow-up study by Sasson and Morrison (2019), where disclosure of the person being autistic improved social outcomes. This leaves a bit of a predicament. For those that do not want to disclose their autism, are they then doomed to be seen as less likable than their peers? What about undiagnosed autistic folks? Do these findings generalize to those with comorbid intellectual disabilities?
I want to preface this article because this is not a direct critique of Unmasking Autism. There are a lot of valuable pieces of advice and information in that book. However, this addresses the combination of factors that led to the idea that neurotypical people hate undisclosed autistic people. Dr. Price’s book does not conclude this, but it has been taken out of context to imply this.
This conclusion has been drawn partly because Dr. Price is a compelling writer who wrote this conclusively without addressing limitations. Part of it is how the studies themselves have presented the evidence. And the most significant contributor is how Tik Tok can spread (mis)information so quickly.
This is common for interpretations of autism studies both on Tik Tok and other media sources. Autism is political in the sciences and as a consequence gets summarized in non-scientific language that may not convey the complete picture of how a study was conducted.
So, let’s break down why we need to take the studies about neurotypical perceptions with a massive grain of salt.
Neurotypical Peers are Less Willing to Interact with Those with Autism based on Thin Slice Judgments
Sasson and colleagues ran three independent experiments, published together in 2017. Different methodologies were used to attempt to generalize the findings along with “natural” interactions of autistic adults and children with researchers. All ASD stimulus participants were “intellectually capable,” according to the study.
The first experiment videotaped 20 autistic adults and 20 “typically developed” adults performing a mock audition for a game show. 214 typically developed undergraduate students then rated the participants on whether they were smart, trustworthy, dominant, awkward, attractive, and likable. They also had “intention” measures of whether the rater would live near the person taped, hang out with, sit next to, and talk to them.
They separated the format that the undergraduate raters watched by transcript, audio-visual, audio only, silent video, and static frame (pictures from the clips). They found that autistic people were rated less favorably in all measures except in trustworthiness, intelligence, and living near (though sitting next to them was also close).
This experiment seems compelling, but there are some key factors that we need to consider here as far as limitations. The first significant factor is that the raters were undergraduate students. It didn’t specify what the undergraduate students majored in, but considering that most recruitment happens through psychology and similar fields, it was most likely that.
This is a huge issue. This doesn’t mean that neurotypical people view autistic people as less likable. It means that a sample of undergraduate students saw autistic people as less likable.
Another issue (which the author acknowledges in limitations) is that the participants were videotaped. This doesn’t translate to real-life social interaction. And this likely had to do (particularly with static frames) with facial features and other atypical physical characteristics of autism, not with social interaction.
Even though the sample size was adequate, the sample itself was flawed and can’t be generalized. The stimulus was also limited. It’s hard to know if the twenty people in the sample were viewed negatively because of autism or if other variables could explain the difference in ratings.
For example, it’s well documented that physical attractiveness affects rater perception. While audio-only and speech content could mitigate this bias, this effect could have greatly affected the other measures. We don’t know anything about the stimulus participants except their diagnosis (or lack thereof). There are too many variables here to conclusively say that autism was why the stimuli were rated poorly, even if there was an association between autism and rating.
The second experiment took 12 autistic adults (2 female, 10 male) and 17 typically developed adults (7 female, 9 male) that had to have a conversation with the experimenter about open-ended get-to-know-you type questions (like “Have you seen any good movies lately?”. They were filmed in POV style with a program that took 10 random pictures through the interaction (excluding blurry photos, etc.).
37 undergraduate students that received credit from participating then rated the frames on measures of “how awkward,” “how approachable,” and “how likely to be friends.” Participants were matched by gender, which is concerning considering there were only 2 autistic female stimuli. The raters were shown multiple clips of the same individuals to test if impressions shifted with repeated exposure.
Raters responded more favorably to the three measures to typically developed individuals, regardless of how often they saw an individual.
Again, these were undergraduate students. We’re examining one sample of undergraduates from one university. They received credit, so we can also presume they were in social sciences. This cannot be generalized.
The researchers acknowledged that repeated exposure does not translate to familiarity. So, knowing someone better may improve scores even if seeing pictures of the same individual didn’t produce a more favorable score.
The problem of visual perception affecting trait rating was especially relevant for this experiment. Again, they didn’t give us any info about the stimuli other than diagnosis, so we have no idea how much physical presentation and related factors played into trait rating. Pictures of interactions do not provide much social information, so concluding that this was evidence of autistic people being rated poorly for social cues is not accurate.
It makes me wonder if we arbitrarily grouped neurotypical people into two groups, would we see similar effects? In other words, how much are individual differences affecting these ratings?
The third experiment consisted of 7 autistic boys and 7 typically developed boys as the stimuli. For this experiment, they had 98 adults and 33 typically developed teens as the raters. The ages of the adults were 19-64, and the teens were 10-16.
Stimulus participants were recorded telling a story of happiness, fear, surprise, and anger. One sentence was taken from those videos where all four emotions were expressed. The boys were rated as likely to start a conversation, have many friends, get along well with others, spend time alone, and have awkwardness. The autistic participants were rated less favorably on all measures by both adults and teens (though not significant on awkwardness for teens).
This sample actually had enough diversity to be decent. The teen sample was not large, but it was recruited through a large sampling database.
I think it’s interesting that they said that autistic individuals were rated less favorably on “spending time alone” (meaning they were seen as spending more time alone). I think that shows a bit of researcher bias as it was implied that this trait was perceived as a negative social trait.
The stimulus sample was the smallest of the three experiments, so individual differences could have played into judgments, similar to the other experiments. The content of speech would have also been very relevant, but the researcher didn’t provide any examples. Also, all stimuli were boys, so this can’t draw any conclusions about how young autistic girls are perceived.
This experiment had some flaws, but less than the other two. Overall, I think this was the best conducted of the three.
Problems with the researcher’s conclusion
One of the factors that went into how Dr. Price wrote their excerpt essentially had to do with how the study was presented.
The researchers presented all three experiments to conclude that neurotypicals create negative first impressions of autistic people, as demonstrated by the various methodologies and samples.
It seems like by using the three studies in conjunction, they were hoping that the limitations of each would be mediated by similar findings in another study.
Considering similar limitations existed for all three (limited stimuli and heavy reliance on visual presentation), and undergraduate students were the raters in the first two studies, concluding that the evidence is “strong” that neurotypical peers view autistic social presentation negatively is not substantiated.
My issues with this conclusion are validated by a follow-up study that Morrison et al. (2019) did that found rater association with autism affected the rating more than traits of the autistic stimuli.
Follow-up study questions their conclusion
The same 20 autistic adult videos from experiment one were used with 505 undergraduate raters. The undergraduate raters were given a series of traits assessments that measured their level of connection to autistic individuals and willingness to interact with autistic individuals the raters had.
They found that among this sample, the higher stigma of autism significantly predicted negative ratings of the videos of autistic participants on all measures except dominance. More stigma was also associated with higher negative ratings when the videos were labeled as autistic.
So, disclosing an autistic diagnosis may only produce a positive outcome if the person the autistic person interacts with doesn’t hold a stigma against the condition. I’m sure this is not surprising to most autistic people, as I do not disclose my diagnosis in every situation (for good reason).
Meanwhile, higher autism knowledge only produced more favorable outcomes when an autism diagnosis was disclosed.
Limitations of this study were the undergraduate sample and not including typically developed videos, so it’s hard to make a direct comparison.
There are many studies to consider when examining how neurotypicals feel when interacting with autistic individuals. There are more studies cited in Dr. Price’s book that I don’t have time to examine in one article. Since the focus is on Sasson et al. (2017) in the video circulating Tik Tok, I felt it was important to break down that particular study.
I also think it is unfair to Dr. Price to take a couple paragraphs from a 300-page book and pick apart the wording of it. That is why this isn’t a review of Unmasking Autism. This directly criticizes this excerpt’s use and the greater trend of misconstruing science in media.
I hope you leave this article feeling slightly less fatalistic about how autism affects your interactions, knowing that the study cited doesn’t apply to all autistic people. It also affirms (and what I think Dr. Price’s greater point was) that autism education is essential to help improve communication between neurotypical people and autistic people.