Teaching Autistic Adolescents to Identify Fear and Anger

Many emotion-recognition programs focus on matching facial expressions to feeling words—but faces don’t always tell the full story. A recent study explored whether autistic adolescents could learn to label fear and anger by focusing on what happens in a situation rather than what a face looks like. This shift matters for clinicians who want to teach emotion language that actually transfers to real life.

What is the research question being asked and why does it matter?

The question was whether autistic adolescents can learn to correctly use “fear” and “anger” by paying attention to what is happening in the situation—not just what a face looks like.

Many emotion programs teach kids to match a face to a word, like “mad” or “scared.” But people don’t always show emotions the same way, and many autistic learners don’t look at faces in the same way. If a learner only memorizes facial pictures, they may not use the skill in real life.

This matters because fear and anger show up in everyday problem-solving, safety, conflict, and coping. If a learner can notice what changed in the environment and what the person is trying to make happen next, they may be better able to talk about feelings and ask for help clearly. This study focused on a specific, teachable piece: labeling the emotion word and describing the “what happens next” part of the situation.

What did the researchers do to answer that question?

The researcher worked with four autistic boys ages 12–13. Sessions happened over Zoom at school, with staff in the room for technology support. Learners watched short animated video clips (about 9–20 seconds) showing different situations. Some clips were examples of fear or anger, and some were non-examples. The videos varied on purpose—different places, characters, items, and even facial expressions—so learners couldn’t rely on one small cue.

The teaching goal had two parts. First, the learner had to say the correct emotion word (fear or anger; synonyms like “scared” or “mad” counted). Second, the learner had to describe the key part of the situation: how “distancing” happened. In their definition, fear meant the person gets away from the harmful thing. Anger meant the harmful thing is made to go away from the person.

Teaching was adjusted using a test–revise–test method. When learners made repeated errors, the researcher changed the teaching steps—like adding clearer context before the problem happens, giving feedback and models, or showing example and non-example clips back-to-back to highlight the difference.

After teaching, learners were tested with new clips they had never seen, and the teacher did not give feedback during these tests.

How you can use this in your day-to-day clinical practice

If you teach emotion labels, consider shifting the target from “matching a face to a word” to “matching a situation to a word.” In this study, learners improved when the program made them track what changed in the environment and what the person tried to make happen next.

In sessions, you can start by teaching emotions using short stories, videos, role-play, or real events. But your prompt should focus on “What happened?” and “What did they do next?” more than “What does their face look like?” This protects against rote responding and better matches the real world, where faces are often unclear.

Use simple, repeatable sentence frames that help the learner say both the feeling word and the reason clearly. The study used a consistent pattern: “He felt scared and wanted to get away,” versus “He felt mad and wanted it to go away.”

In practice, keep the frame short and natural. Accept the learner’s own words as long as the meaning is correct. The key is that the learner connects the emotion word to the function of the behavior in that moment—moving away versus making the thing leave. This isn’t about sounding polite. It’s about helping them describe what they need.

Plan examples and “close” non-examples so the learner learns the boundary of the concept. A close non-example changes only one important feature. Here, fear and anger were close because both involved something harmful, but the direction of distancing changed.

If your learner calls everything “mad,” build teaching sets where the only difference is whether the character runs away (fear) or pushes/throws/chases the thing away (anger). If your learner calls everything “scared,” do the same. This kind of tight comparison reduces overgeneralizing.

When learners struggle, treat errors as information about stimulus control—not as noncompliance. In this study, common errors were labeling objects (“cupcake,” “controller”) or saying an action without direction (“running,” “walking”). That tells you the learner may not be attending to the critical feature you care about.

You can respond by prompting direction words (“away from” vs. “go away”), adding visual cues (arrows, icons), or asking a more specific question (“Did he move away, or did the thing move away?”). Make one change at a time so you can see what actually helps.

Consider using juxtaposition on purpose. When the researcher placed an example right next to a matched non-example—same setting, same characters, different outcome—some learners improved. You can do this with pictures, short clips, or quick acted skits. Show Scenario A and Scenario B back-to-back, then ask what’s different. This supports discrimination without extra rewards or long explanations.

Keep assent and engagement as part of the plan, not an afterthought. In the study, learners could pause or end sessions, and some needed shorter sessions across more days.

If the learner walks away, argues, or shuts down, that’s a signal to adjust the task or pacing. Offer choices like “one more video or break,” let the learner control the pace, and reduce distractions. Emotion work can feel personal, confusing, or tiring—you want the learner to stay safe and willing.

Be careful about overgeneralizing what this study shows. It included only four learners, all cisgender male adolescents, and all teaching happened through Zoom animations. This doesn’t tell you the same method will work for all ages, learners with limited speech, or in real-life peer conflict. It also doesn’t show that learners could label their own emotions during actual upsetting events.

Treat this as an early, structured way to teach one piece of emotion language. Then test whether it helps your specific learner in natural settings.

Finally, don’t assume your definition of “fear” or “anger” matches the learner’s lived experience. Use this approach as a starting point, then validate it against the learner’s real behavior and goals. The best clinical use is to help a learner describe what is happening and what support they want—in a way that increases choice, safety, and dignity.

Works Cited

Linnehan, A. M. (2025). Teaching autistic adolescents to identify fear and anger: A preliminary study. Behavior Analysis in Practice. https://doi.org/10.1007/s40617-025-01129-x