Is it always easier to talk with people who are like you?

If we’re both autistic, will we automatically get on better?

Not necessarily. Shared experiences can make some things easier to explain. But whether two people click also depends on what they want to talk about, how they express themselves, how well they know each other and whether they both have the energy to keep going. Research comparing autistic and non-autistic communication shows group results on a specific task. It can’t tell you who you should spend time with.

In these studies, “the same neurotype” just means people were grouped as autistic or non-autistic. It doesn’t mean everyone in a group has the same personality, needs or way of communicating. Keeping that in mind stops “I’ve finally found someone who gets me” from turning into “everyone with this label will get me.”

Section sources [1] [2]

How did the experiment measure “successful communication”?

In 2020, Crompton and colleagues recruited 72 adults and split them into nine chains of eight: three all-autistic, three all-non-autistic and three that alternated between the two. A researcher told the first person a story containing 30 details, and each person passed it on to the next. Each retelling was scored on how many details survived. Separately, participants rated how easy and enjoyable the interaction felt.

Think of it like the telephone game. To give a made-up example: the original message includes “Saturday, two o’clock, south entrance.” If one person drops “south entrance,” the next person can’t pass it on however carefully they listen. So what’s left further down the chain depends partly on that one exchange and partly on what reached them in the first place. And a score on this task says nothing about someone’s overall social ability.

Section sources [1]

Where did the two studies agree, and where did they differ?

In the 2020 study, the mixed chains lost more information, while the all-autistic and all-non-autistic chains did about the same. The 2025 follow-up, with 311 participants, again found no difference between the two single-type chains — but it didn’t find that mixed chains did worse overall. So you can’t conclude that autistic people can’t communicate, and you can’t combine the two studies into “it’s always better with your own kind.”

Statistics and personal experience are different things. A study finding no difference between groups doesn’t mean the misunderstandings and tiredness you feel aren’t real. A study finding an average difference doesn’t mean every pair of people fits it. If you’re comfortable with a particular friend, that matters in its own right. It doesn’t need a study to back it up.

Our reading is this: both studies show you can’t assume someone is bad at getting information across just because of their diagnosis. But being “the same type” doesn’t explain every interaction either. The diagram below simply sets the two studies’ results side by side. It isn’t a ranking of anyone’s social ability.

Section sources [1] [2]

What changed in the second study, and does that explain the difference?

The 2025 study ran at three sites in the UK and US. It used chains of six, and each chain passed on both a made-up story and a factual passage. The first person got the material from a video. It also compared what happened when people were or weren’t told about others’ diagnoses. The 2020 study used chains of eight, one made-up story told in person by a researcher, and participants knew which kind of chain they were in.

The authors discussed how a more varied sample and the method changes might have mattered, but didn’t pin down which change caused the different result. They used chains of six so that the story wouldn’t run out of details to lose towards the end. Since the 2020 study had already shown differences within the first six people, they thought the shorter chains probably weren’t the whole story. For readers, “the results were different” and “we know why they were different” are not the same thing.

Here’s a comparison: if you change the ingredients, the pan and the heat all at once and the dish tastes different, you can’t say it was the pan. Research measures things far more carefully than cooking does, but to know which change made the difference, you still need evidence that separates it from everything else. Don’t pick the explanation that sounds nicest and treat what nobody knows yet as settled.

Section sources [2]

If the message got through, why is it still so tiring?

You might get the time and place of a party exactly right but spend the whole evening straining to keep up, and come home shattered. Or you might have a lovely chat and get the time wrong. Accuracy, comfort and whether you want to meet again are three separate questions. Real conversations also involve asking questions as you go, rephrasing and clearing up misunderstandings — none of which a one-way relay can capture.

Chapter 8 of The Complete Guide to Asperger’s Syndrome talks about getting a stalled conversation back on track. Not knowing how to say “I didn’t follow that,” or how to explain something a different way, can bring a conversation to a halt. That’s worth paying attention to. But the book is older and sometimes generalises about everyone, and we don’t follow it there, nor do we use it to explain these studies’ results.

The “double empathy problem” is a reminder that people with different experiences can misunderstand each other both ways, so communication problems shouldn’t always be blamed on one side. But “both ways” doesn’t mean every disagreement is split fifty-fifty. Who misunderstood what, and who can change what, still depends on what actually happened that time.

Section sources [1] [2] [3]

Can these findings be applied directly to my child or to me?

Both studies involved adults doing a spoken English task. They weren’t classroom interventions for children. The 2025 authors also say it’s unknown whether the results would hold in other cultures, for people with learning disabilities, or for people who don’t communicate through speech. So these studies can’t decide which group your child should be in or which training they should do — and they certainly can’t show that someone who doesn’t speak can’t communicate.

That doesn’t make the research useless. It’s best at helping you question assumptions that are too simple, like “if they can’t explain clearly, it must be their fault.” It isn’t an answer sheet that fits everyone. What support someone needs comes down to whether the message actually got through, whether their way of communicating was respected, and whether both people can keep the conversation going.

Section sources [2]

How can you use this in everyday conversation?

Work out which step actually went wrong. Here’s a made-up example: a friend says, “We’ll go when you’re ready.” You think that means grabbing your bag; they’re waiting for you to change your shoes too. You could ask, “What does ‘ready’ include?” They could also spell it out: “Get your bag, put your shoes on and meet me by the door.” That way the listener can check, and the speaker fills in what they’d left out.

If the message is clear but talking is still exhausting, talk about pace, surroundings and taking breaks, rather than checking over and over how much you remembered. Choose friends based on what being with them is actually like: can you disagree, will they explain when you don’t understand, can you stop when you’re tired? Shared experience can be a good starting point. But you don’t have to force a friendship because you share a diagnosis, and you don’t have to rule out people who are different from you.