Academics, write clearer papers!

We seem to be at it again, taking reasonable work by reasonably naive people, taking their extreme estimates, and gaslighting it as their only prediction on social media as clickbait for angry people.

I attempted, despite thinking he was a dilbert, in this to defend Ferguson against lynch mobs last year over this, and we seem to be doing it all over again about Warwick’s prediction.

Here is Warwick’s efforts, and everyone seems to have got very carried away with the fact that this plot in fig1 on page3 seems to suggest we might hit 5k/dy deaths in Jul.

Leading to many frustrated narratives like this from Snowdon on his blog. Now I actually have a lot of time for the man, standing up to @fatemporer and @toadmeister makes him a hero in my book, but I’m worried he’s fallen into the same gaslighting trap of misquoting people for their efforts that Toby did over Ferguson. In Toby’s case it is hard to tell if it is duplicitous, in Chris’, I think it is mostly down to how badly written Warwick’s paper is.

I review statistical research everday, and even I am having to look for truth in their effort. I’d say it’s even worse than Ferguson’s effort, and what these academics don’t realise is the world has changed, you will be judged by lay people on your efforts, so buck up your ideas, don’t lay things out in unclear order, and written in academic wank speak. Try harder.

Specifically, the interpretation mistake being made about the plot above is that it makes these assumptions:

• vaccine rate of 1m/wk
• low efficacy, 2/3 of the trial results
• the Jan2021 lockdown limiting deaths to a 800/dy

But, this does not match our reality, which is more like:

• a vaccine rate of 3m/wk
• high efficacy, better than the trial
• the Jan 2021 lockdown hitting 1360/dy deaths

Now they do model that, on page4 in fig2 they cover scenarios with Jan peaks of about 1.3–1.5k, here they are.

Specifically the purple lines covering 3m/wk vaccine rollout and high efficacy. In these scenarios they forecast between 0 and 400/dy deaths.

Which actually matches all of our expectations.

We need to guard against a Toby-style gaslighting, as I summarised in this tweet about this article, the scenario in Ferguson’s famous Mar16 paper that most closely matches out current scenario estimated 120k deaths. Let’s not gaslight naive researchers who haven’t realised they really need to up their game in writing papers that will be read by the masses… although, seriously, universities really do need to prep academics better for writing for the real world.