Less NeuroRant, more AcademicsNotUnderstandingVariabilityRant
So, in the pit of hell where the NeuroRant idles the days away, it is sacking season. Senior academics have got their hunting licenses all in order (approved by human resources no less), and are looking for prey.
So who do they go for? Well obviously those who have failed to publish enough and get big grants, obviously. Why them? Well because, obviously, they are dead wood, who don’t have what it takes to be superstar world beating academics. They are lazy. They are stupid. If only they could be quietly pushed away onto a slag heap of poverty and unhappiness, then they could be replaced by shiny new academics who are energetic and sparkling, and will usher in a new era or plenty. This time everything will be different.
Everyone agrees with this approach. Those who don’t agree with this approach are soft, or don’t get enough grants and papers and so are scared. It’s academic Darwinism. It’s great.
Unfortunately it’s wrong. There are lots of reasons why it’s wrong, but I’m going to focus on one big one: senior academics have no understanding of randomness. They see only signal and no noise. Simply put, they don’t understand the regression to the mean.
There is random variability everywhere: especially with getting papers and grants accepted (due to randomness of reviewers, editors, being beaten to publication by another group etc etc). Taking a step back, there is variability in getting results from experiments that make sense or are what you predicted, randomness in a new technique working, in an MRI machine not breaking down; all of which makes getting grants and papers vary randomly. I don’t mean it’s all random, but a big chunk of it is.
This means that even if all academics are equally shiny and energetic, some will randomly fail to get grants and papers, and so be in the dead wood category. The same thing happens with football managers (read about it here).
So, many academics are sacked who are as good (in terms of future grants and papers) as the people that will replace them. And as with football managers, even if we ignore the fact that the punishment has a large arbitrary, unfair aspect to it, sacking people comes with costs. In academia, all those start-up costs for new staff, all that disruption to research and teaching programmes, fear and stress on non-sacked academics, arguments about space for labs, not to mention the costs in human resources time for delicately sacking someone (poor human resources). Therefore, if people are sacked because of regression to the mean effects, it’s a negative sum game for the university. It’s just stupid.
But, what I find interesting, is that it’s not just senior academics who don’t understand the regression to the mean, it’s also the other academics who aren’t being sacked, the successful academics. Those who are getting grants and papers with apparent ease. In fact, they are actually enjoying the other side of the regression to the mean. Their success is also, in part, the result of random variability and it probably won’t keep going. But those academics also ignore this randomness, and put it all down to their own brilliance, while agreeing that the dead wood needs to go.
And so the system trundles on. All these pompous academics beating some of their members up and falling over each other to fawn over others, unaware that this whole pecking order is fundamentally flawed. Like a bunch of stupid, fighting penguins or something.
There’s got to be a better way. Maybe we could try to be nice and support each other for a change.
Disclaimer (NeuroRant is judged successful by their institution’s standards if not their own).