Thank you! It’s been a long process but honestly, pretty proud of our approach. Not a type of research I do anymore really, but the rigour is something I’ve enjoyed doing and now showcasing! Thank you!

(As many of you know I've also moved away from these methods these days, this paper goes into detail why in the discussion)

The biggest thanks to Jess and Alex - and QJEP! - for being on this science adventure with me!

Amazing to be in a position where we don't contribute to the 'file drawer' problem - we tested a reasonable question and found the best answer to it. Just because it was a non-effect does not mean we shouldn't share it.

What did we find in the end? The largest, most robust study, failed to replicate the effect of interest in study one. We found some other things in the study but the more interesting thing was being able to test a reasonable hypotheses and then fail to replicate it in a journal space.

For the RegReport we sent studies One and Two for review and said we will conduct a third (with a large sample size conducted at both UWin and again DMU). Regardless of what the results are, will you publish this robust test? They agreed! (I’m a big RegRep fan!)

We decided to do a tie-breaker study. I was then at the University of Winchester and access to a new sample. With Alex Jones' (Swansea) Bayesianising and analytical lens we pitched a Stage One Registered Report to QJEP.

The findings didn't replicate (see paper for all the ways we tested this). So what do we do? Maybe it's a South/Midlands divide? Maybe demographics of the University? Or maybe it just isn't a robust finding?

A few years later I had the chance to meet Jess Hall (DMU) and began to collaborate. She agreed to run a replication study at DMU in Leicester. Same stimuli, same equipment, but we had the chance to expand our sample and stimulus set. So we went for it!

But this study had a pretty small sample size and limited stimulus people (only female targets). I wasn't confident to publish this yet.

I found what I'd hoped - when people thought someone was more threatening they spent less time looking at a face, but spent time looking at the whole body (I was doing a PhD on the importance of studying walking in social perception so kind of liked this!)

During my PhD, I conducted a study where we tracked eye movements of participants looking at people walking. I was interested in how social perception affected our gaze behaviour - do we look more at faces or bodies if we're more threatened? (A reasonable hypothesis based on literature)

This our new Stage Two Registered Report twice internal failure to replicate...

What does that all mean - for both this study and doing science?

As we note at the end - people should trust these findings exactly as much as they should trust ‘social media addiction’ scales ‘validated’ in the same ways 👀 (as in, please don’t 😂)

Yes! Our choices were to follow exactly what others do when designing the same tools to make our point about current approaches (we have the reviews on the appropriateness 😃).