rockym93 dot net

Master's Thesis

I spent most of 2017 doing research on anthropomorphism in science writing, and how it affected the way people thought and felt. If you're incredibly keen, you can read the full thesis here, my extended literature review here, and my conference slides here. If you're less keen (and I don't blame you, it is very long), I've summarised what I learned below.

What is anthropomorphism?

Anthropomorphism is the psychological phenomenon of giving human traits to things which aren't human. It happens to everyone, in pretty much every culture throughout history. Anthropomorphism is why we give our pets personalities, why we name our cars, and why we make friends with volleyballs when we're stranded on desert islands.

Our brains do this because it helps them feel like they understand the world better, and because we're social animals that rely on our relationships with other humans to survive. That means we're particularly likely to see things as human when we're confused or lonely.

Why do we care?

Well, I care because I was told pretty explicitly not to at one point, but when I dug into it a little deeper I realised there was no clear answer as to why.

You care because anthropomorphism is built into the way we think, write, and tell stories - and that includes stories about science. There are some strong opinions that anthropomorphism is unscientific, it causes mistakes in thinking, and should be considered a misconception. But there are also tantalising hints that it can help us tell more engaging, empathetic stories. People have studied how it affects kids in classrooms, but not how it affects adults who are choosing to read about science in their own time.

My research tried to find some evidence for these ideas.

How did we find out?

I asked 174 strangers from the internet, split into three groups, to answer some questions.

One group read a blog post starring a plant that acts like it's a person - Diary of a Space Zucchini by NASA astronaut Don Pettit. Another group read the same post, but edited to be totally impersonal. A third group read nothing at all, and just answered the questions for comparison.

plant vs plant

I looked at the answers to those questions for evidence of the kinds of misconceptions and mistakes that anthropomorphism is supposed to cause. I also looked at how much each group enjoyed and connected with each piece of writing.

What did we learn?

The effects of anthropomorphism are pretty small. It made no difference at all in the frequency of mistakes people made, or in how connected or engaged they were - although both groups which read a post did much better than the group that read nothing at all.

correct answers, by group

We saw some evidence that people who read the anthropomorphic text had stronger ideas about plants having feelings, and stronger emotional connections to the plant in the story. These ideas weren't necessarily seen more often, but when they were, they were more strongly expressed.

People who read the anthropomorphic text also tended to explain their answers slightly more. They said things like 'I don't know' and 'just because' - things which didn't actually answer the question - significantly less often.

non-answers, by group

However, they said they felt slightly less confident with the content. Still pretty confident, mind you, but less so than the group with the straightforward text.

percieved competence, by group

And that's about it. We found no other statistically significant differences, although there was loads of fascinating stuff to dig through in the answers people gave - but you'll have to read the full thesis if you're interested in those.

So what?

So, anthropomorphism doesn't seem to be any different from any other metaphor or example. It might make writing more memorable - but it might also become a little distracting. However, it doesn't cause the kinds of wildly inaccurate assumptions which everyone was worried about.

It's probably best to steer clear of it where it's not clearly a metaphor. Things like explaining animal behaviour, or discussing evolution (especially in the context of creationism), where pretending something is human might genuinely be confusing.

In general though, you're probably okay to use it in your science writing. Just make sure it's helping to illustrate your point, and not distracting from it. Use your judgement. Stick to your message, and know and trust your audience, and everything will be fine.

space plant!