Cynical Archeology
You find some odd stuff cleaning up your Dropbox
Most of what I’ve written is in some poorly curated folder, somewhere. The act of writing, for such a long time, was simply curative — some exegesis of some feeling (usually anger) or doubt (usually scientific). That was all. Letting people read it was an afterthought. Maybe I never thought they’d care.
Most of it, also, is incomplete.
So it’s somewhat unusual to stumble across something I finished but also never released, but I have.
This was written in 2013, pre-supposing a great deal of the conversations we’ve had about research problems and analytical flexibility in the meantime.
Previously, I’ve tried to pinpoint the beginning of my … methodological dissatisfactions. This piece helps — it shows some kind of (hacky but mostly complete) awareness of over-narrativisation, p-hacking, HARKing, outlier fiddling, crappy press releases, the paper-to-book pipeline, WEIRD problems, and bad design decisions. I probably didn’t know these things had names back then.
I should note I don’t think this knowledge particularly unusual. I don’t think I’m displaying a great deal of foresight here. This is a very old road.
But I suppose if there’s one point, behind the fact that you might enjoy my previous bleakness, is the fact that the below seems a bit tired now. Well-trodden ground.
That means things have changed.
They’ll change more.
HOW TO BE GOOD AT POPULAR SCIENCE
By James Heathers
You’re so very educated. You’ve done a doctorate in the social sciences, and you make everyone call you doctor because you suffer from the nagging suspicion that you don’t actually know anything. Have no fear, you’re starting your own research program now. Humanity needs Help, and you’re here to understand it. Now, how should you do that?
You need to start with an Idea. Not just any idea will do for your Idea. It needs to have spice, some kind of visceral appeal — love, bonding, violence, sex, intimacy — or adorable animals (highly encouraged), or at the very least some gosh-darned gorgeous heatmaps of a brain.
Pay no mind to anything revelatory, some searing insight which is easily demonstrated by an experiment but oh-so-hard to find come across. You may never find something so straightforward. Avoid insight; it requires a lot of foolish reading and late nights. Your idea, rather, should be Cute. Not clever. Clever is for scientists who have things in Tubes.
Your Idea should also be flirting with counterinuition — not completely pedestrian, not outlandish, but oblique. It’s the first thing on the tip of your tongue first thing after you offer a normal explanation “ … or maybe it could be… ?” “Actually, it turns out that…”
Just novel enough to be unusual.
Just believable enough to be pulled into the orbit of convention.
Figure out your model. This is not a Model, this is a model; the details are not important. You need large concepts to knock up against each other, Big Important things. Depression, approach, withdrawal, attachment, self-control, hatred, self-hatred, control of self-hatred, the entire nervous system, the concept of love itself…
Stack up these ideas like so much firewood, and don’t hold back. If you need to measure anxiety, ask about it at the time — “How anxious are you right now?”, and in the past — “How anxious were you?”, and normatively — “How anxious are you normally?”, and non-normatively — “How anxious are you today in particular?” Split these into sub-categories. “Were you anxious? Did you have a premonition? Were you catastrophising? Were you worried? Concerned? Are you concerned about how often you worry?”
Never mind if you measure the same thing more than once. This is you being Careful. What if you don’t quite get the concept right the first time?
Never doubt how people tell you they feel. Asking a 19-year old undergraduate at a sterile workstation answering an online questionnaire is a good predictor of their hypothetical violent capacity in 15 years after they lose their second job. People know how they feel. Why wouldn’t they?
Hammer these concepts crudely into place, then get a big gaggle of young, white people, and leave them unsupervised to give you answers. Do not worry about the pesky details of data collection, just acquire numbers which map onto the measures you’ve chosen… sort of. Push it all together in an analysable form…. sort of.
And now you have your data. It won’t make a ha’penny of sense, of course, because it is the conceptual equivalent of a ball of yarn shared between two hyperactive kittens.
What you need to do is nudge it. Some of those measurements you made will be in the wrong place, of course. Just prod them a little. Some of those questionnaires won’t work well. Forget them, they obviously weren’t meaningful. Some of the values will be a little strange — they obviously aren’t right. Quietly retire them. Maybe you need a more complicated statistical model. Add some layers, some covariates, some mediators, some strange attractors, some balance, some cumin, some normalisation, or anything else that sounds like it is Right and Proper.
(Do this *just* enough to maintain your veneer of respectability, but not enough to attract the outraged attention of people with a lot of nosy impertinent questions.)
This is where the Narrative lies. The Narrative is what you need. A story. A franchise. A concept which connects the ideas you’ve got left behind. They all relate to each other somehow — figure out how. It’s not foreshadowing, it’s emotional pre-foreshadowing. It’s sub-social exclusion perception. It’s pre/post-frontal disengagement. Relish in these happy, shiny words. They Help a great deal.
Now, write.
Polish every sentence like you’d curry a pony. Make every sentence a glowing beacon of humble straightforward insight. Paper over all the cracks in this raw, messy, grasping process. Plan all your seat-of-the-pants analyses in retrospect. Fulfill expectations, soothe doubts, pose a few cheeky questions, position, angle, finesse. Imply you know more. Hint — never say — that you may have discovered a major law of human behaviour. Save that for the Press Release.
You will write one of those, of course. Feel free now to be a little more mellifluous, set your imagination on a longer chain. Your humble-brags will quickly be upgraded by staff reporters to ‘a genius scientist solves cancer, athlete’s foot and the number 4' anyway.
Don’t correct them if they take it too far — they mean well. They’re Helping.
Other journalists will call. Be nice to them. Help them to see.
Run this through a few times, and you are famous. Your likeness will be printed on the Cover of Things. People will pay to hear you speak. Your theory is lovely. You’re doing so well. It is time for Your Popular Book.
Your Popular Book should embrace your research — which by now, of course, is a foregone conclusion — and be quite short, well meaning and slyly expensive. Your Popular Book is your passport to freedom.
At no point, under any circumstances, should you entertain the idea that you are doing anything less than the right thing. Science is conducted for the benefit of humankind, not for naked personal gain, and you are Helping.
