Taking the “Fear, Anger and Joy” Test

This week for Take-a-Test Tuesday, I took the “Fear, Anger, Joy” test. It’s not an Asperger’s test, but it does measure something that aspies often struggle with: reading emotions through facial expressions. Also, since the last two tests have been question-oriented, I thought a visually-oriented test might be a nice break.

testmybrain.org is a citizen science project run by Harvard University. The website is used to conduct large-scale behavioral research studies with the goal of developing neurocognitive tests. By taking one (or more) of the tests on the website, you contribute to ongoing research and you also get a nice little write-up about your results, what they mean and why the research is being done.

I like the interactivity of the tests and the novel approach that Test My Brain takes. For example, the test that I’m taking today, “Fear, Anger and Joy,” is a test of the ability to identify emotions in facial expressions. Unlike some of the other facial expression recognition tests out there that allow you to study an expression for an unlimited time before identifying it, this test only allows you to see the facial expressions for a couple of seconds. I feel like this is closer to what happens in social situations, where facial expressions are fleeting.

Pros and Cons of Fear, Anger, Joy


  • Automated and self-scoring
  • Better simulates social situations by limiting viewing duration for each set of faces
  • Detailed explanation of scores
  • Results include average scores and percentile ranking so you know how you measure against others who’ve taken the test
  • Explanation of why the research is being done with a link to more info
  • Taking the test contributes to ongoing research


  • Test is part of active research so no conclusive results
  • No direct link to AS/autism/autistic traits
  • Forced choice between 2 options equals a 50% chance of being correct when guessing

Taking the Test

This test takes 15 to 20 minutes so be sure you have the time to complete it before you begin. To take the test, go to testmybrain.org and click on the Go! button next to the “Fear, Anger, Joy” test (you may have to scroll down to find it–the order of the tests changes).

You’ll be asked to make your browser window large. I kept mine the size I normally use and it worked fine. The next screen is a simplified informed consent form. You’ll be told what the research is being used for and asked to consent (agree). Once you do, you’re officially a research subject! The next screen collects some demographic information. It’s all anonymous–you won’t be asked any identifying information to get your results or anything annoying like that.

Once you’ve entered your demographic info, you’ll begin the test. There are a couple of practice questions before the first set of faces appears. You’ll see two faces side by side and be asked to identify which is happier, sadder or angrier. There are 56 sets of faces for each emotion. That sounds like a lot until you realize that each set is visible for only a couple of seconds.

After you’ve completed all three sets of faces, you’ll be asked if you encountered any technical problems or if you cheated in any way. Researchers think of everything. Once you confirm that your results aren’t tainted, you’ll be taken to the scoring page.

Scoring the Test

The aim of the test is to evaluate how well people can detect subtle differences in the level of emotion conveyed by an expression. I found the expressions presented at the beginning of each to be obviously different. One face was clearly much happier, sadder or angrier.

As each section progressed, the differences in the expressions became much less obvious and I found myself guessing more frequently. I also realized that scared and happy can look similar to me. The same is true for scared and angry.

Here are my scores (out of a possible 56 on each section):

Which face is angrier?: Your score was 45. The average score is 45.18.You scored higher than three out of every ten people who took this test.

My rankings on the anger section
I scored slightly below average on the anger section.

Which face is happier?: Your score was 54. The average score is 49.49.You scored higher than seven out of every ten people who took this test.

Nearly a perfect score on the happy faces!
Nearly a perfect score on the happy faces!

Which face is more afraid?: Your score was 47. The average score is 43.18.You scored higher than six out of every ten people who took this test.

And a little bit above average on the afraid faces.
And a bit above average on the afraid faces.

I did better than I expected to. I’m not surprised that I score the highest on the happy faces because happiness is (for me) most clearly expressed around the mouth (smiling) and that’s where I most often focus my gaze.

Anger forced me to look at the area around the eyes more and I had to purposely do that. For fear, I quickly found a pattern that helped me: furrowed brows or foreheads. I spent the rest of the “afraid” section looking first (and sometimes only) at the foreheads of both people to see which looked more scrunched up. Maybe I should have confessed to cheating?

Apparently I have trouble taking in a facial expression as a unified whole. I seem to look at the various parts for clues and then try to sum them up.

I also found my perseverative nature kick in when I had to change from finding the angrier face to finding the happier face. On the first of the happy practice sets, I chose the angrier face because that’s what I’d been doing for the previous 58 trials. Oops.

The Bottom Line

This test was fun to take and I felt like I learned something about how I process facial expressions.

How did you do? Were you surprised by the results?

33 thoughts on “Taking the “Fear, Anger and Joy” Test”

  1. I don’t know if that’s the cheating. Maybe you did it consciously while NTs do it subconsciously, but I think NTs probably look for the same cues that you mentioned when deciding whether a face is angry, sad, or happy.

  2. Heh, another well researched post:-)

    I took the test. It was fun, but very difficult due to the high speed. I didn’t fare very well, I had:

    – Which face is angrier: 38 (‘You scored higher than one out of every ten people who took this test’)

    – Which face is happier: 43 (‘You scored higher than zero out of every ten people who took this test’)

    – Which face is more afraid: 40 (‘You scored higher than two out of every ten people who took this test’)

    HOWEVER, it was because of the high speed, not because I can’t tell face expressions apart. With some of them, I didn’t have time to look at both and make up my mind, so I had to guess. In some cases I pushed the wrong button accidentally because I couldn’t control my fingers that fast, knowing in the same second I clicked that it was the wrong choice… but I didn’t have time to stop the move! This is really a racer test! I don’t think old people would be able to keep up!

    Some of the faces had identical expressions (and, I think, were the one same photo), but I could definitely tell the difference with most of them and would have done much better on the test given slightly more time to each.

    Forced choice between 2 options equals a 50% chance of being correct when guessing

    I can see that, but I have a suspicion they have factored that in. Many of the photos are showed over and over again and I imagine there is some sort of algorithm that detects if there is a random choice pattern (or maybe that is just my imagination, I am too lazy/don’t have the brains to think it through).

    1. I had the same problem with occasionally pressing a key other than one I intended to. The speed of the test flustered me and forced me to guess more often than I would have if it went at a slower pace. I was also wondering if a few were the same photo or if perhaps they had photoshopped one very minor change into one of two identical photos.

      When I pointed out the forced choice issue, I was thinking about how a “don’t know” option might alter the results. I would have chosen it on probably 30% of the faces for each set, but instead I had to guess. I’m not sure if they would be able to statistically correct for guessing, though probably must factor into it somehow.

      1. I think they want to force people to guess because even when you feel you are guessing, you may subconsciously be processing some information – maybe enough to make a correct choice.

      2. Yes I think that is their intention.

        There’s an unintended bias though: it is not enough to know the right answer, you also have to push the right button very snappy and then move on. So it unintentionally tests people’s hand-eye coordination speed, and a poor result can stem from not being able to move the fingers fast enough. I imagine that for example elderly people would perform poorly on the test for that reason; thus pull down the average.
        (and me:-)

  3. I scored really high on all these categories, with the fear test being the highest of all of them. I thought this was odd. I’m nor sure what to make of it, as I don’t do well in real life or with other tests on facial expressions.

    1. I don’t know what to make of my high scores either. One of the differences between this test and real life for me is that during the test had to look at the faces, which I often don’t do in RL social situations.

      1. I scored below average for everything.. way below.. me.. the intuitive, people loving, social sciences, empathic to the point of mental illness human and my autistic partner who says “ignore my body language, it’s wrong and I can’t read yours” to people scored above average for everything…

        1. This test is confusing. The autistic folks are knocking it out of the park with above average scores (someone on Tumblr posted 2 out of 3 perfect scores), which makes no sense.

          Did you feel like you were having trouble with the expressions or did you feel pretty confident as you were taking the test?

      2. I knew I was putting the wrong answers in but I couldn’t process in time for the test. My partner and I are just discussing the decoding strategies we both use and the contextual information not included in the test…

  4. Interesting test. I scored high on all three as well and did not expect that. My highest was in the happier section. I get very confused with real life faces, but my theory is that I can read people’s “true” emotions. I think their faces are hiding the real emotions so it causes me confusion because what I feel and see do not match. However, with these still images and no body language or verbal interaction I have nothing to jumble up what I see in their face. ??

    1. That’s an interesting point. I think I read people’s tone of voice more so than expressions to pick on emotions. Perhaps this is part of how you’re reading the true emotions vs. the expressions in real life?

      1. That would make more sense,picking up on their tones. I am not usually looking at their face, my eyes dart all over the place then, land on their forehead, eyebrows, or I try to read lips because I cannot process what they are saying. Faces make me uncomfortable. I really haven’t a clue. I just know something seems off, but I do not know why. 🙂

        1. I rely somewhat on lip reading as well. I didn’t realize how much until recently when I was watching a show on Neftlix and the sound was off slightly from the video because the streaming was messed up. It was driving me crazy the way the actors’ mouths were a bit ahead of the dialogue and my husband didn’t even notice!

  5. I am surprised that you guys who are all diagnosed as Aspies scored high on the test, while I who am not and who identify as not having any particular trouble with reading face expressions, scored rock bottom.
    As mentioned in the comment above, the speedy finger coordination that was required to select the right options fast enough, was difficult for me. So was the fast decision making required. Was no one else troubled by the aspects of the test that were not directly about face-reading?

  6. Wow I got practically 100% on angry faces, ‘scored higher than 9 out of every 10 people’!

    Dead in the middle of fear but only 2 out of every 10 on happiness though!

    It seems like this is another ‘spiky profile’ issue, 20th percentile vs 90th 🙂

    They should do boredom, I know for a fact I can’t tell when people are bored of what I’m talking about 😉 …I suppose that would be subtle signs of unhappiness really, so my lowest score being there would make sense 🙂

      1. It depends on how ‘polite’ the other person’s being, also how excited I am by what I’m saying. Also averted eye gaze probably makes me think they’re trying to listen harder 🙂

      2. Boredom or disinterested is difficult when you rely on verbal tone and pace to gather mood data. NT’s rarely say anything when they are becoming disinterested out of politeness. I think

  7. I wonder whether people scoring highly on the ability to judge emotion in a split second might relate to the Intense World Theory, that we have too much empathy and find other people’s emotions too intense. Maybe a split second is more manageable? http://www.wrongplanet.net/article419.html

    I also think people on the spectrum are way more likely to have been bullied in childhood, so more likely to have obsessed over reading danger signs in the faces of others.

    It would be interesting to see how the scores differ when there’s no time limit to looking at faces. How do people compare on the ‘Mind In the Eyes’ test? http://glennrowe.net/baroncohen/faces/eyestest.aspx

    1. I get around 24 to 27 for the eyes test depending on when I take it, I mostly feel like I’m guessing but when I can do it, it’s by imitating the facial expressions and seeing how it makes me feel.

    2. It may be the case that seeing the expressions so briefly is less stressful. I’ve read about the Intense World Theory and find I can relate to much but not all of it. I lean more toward the issue being emotional dysregulation (having an on/off switch with no volume control) rather than too much empathy. Although empathy is a tricky word and those two things may be one in the same, depending on a person’s definition of empathy.

      I wrote up the Mind in the Eyes test and I’ll post it next Tuesday so people can compare their scores. One thing that struck as I was researching this morning is the difference between reading simple emotion (fear, anger, happiness) and reading complex mental states (theory of mind). So yes, it would be every interesting to compare scores on the two. I got a 31 on Reading the Mind in the Eyes this morning but I’ve taken it before so my score may be biased.

      Also, I’m going to post the AQ-10. I have it written up already so perhaps a week from Tuesday on that one.

  8. As a girl with Asperger, I usually have problem telling emotions on the face of other people, especially if i’m not familiar with them, unless it’s really apparent. But I scored high on this test, which surprised me! For anger and joy, I had no problem, but the fear was tricky. When i think about it, I was just guessing everything over the form of the mouth and eyebrows…
    Nevermind, those tests are funny (even if that’s not the goal, I still find it funny)
    Sorry if there is mistake, english isn’t my first language!

    1. Lots of aspies/autistics scored high on this test. It’s puzzling! I also found myself focusing on the mouth and other parts of the face too, especially for the anger and happiness parts. I wonder if fear was harder for us because it’s expressed less clearly in the parts of the face we tend to look at?

      Also, your English came across fine. I hope you’ll be a regular commenter. 🙂

  9. They’ve changed the test to “Your Social Brain” and it involves a second component called “Social Perspective Taking” in which you “…were presented with a series of stories that described people in a variety of unusual social situations. This test was designed to test your logical reasoning and to see how you were able to follow a story that had multiple perspectives. Your score represents how well you were able to remember and evaluate complex stories.”

    There no longer seems to be a breakdown of the different faces but an overall score for both sections. For instance, I’m not diagnosed with AS but I strongly suspect I have it. I scored 87 overall for the face recognition. This could be due to not being timed and you have to pick out 3 or 4 faces from a group of 12. I scored higher than 6 out of 10 people. For the second part, the Social Perspective Taking I scored 15/15 and better than 9 out of 10 people. I have no idea what relevance this could have to AS though.

Share Your Thoughts

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s