The research on this is not new, and it’s not as simple as this. If you give people a task that’s well-rehearsed for them, that they’re good at, they’ll tend to do better. But if you give them something less well-rehearsed, something that’s “hard,” they do worse. I wish I had my psychology textbooks on hand for the references, but a lot of this stuff isn’t obvious until after-the-fact.
The full name is the "Yerkes-Dodson Law of Social Facilitation." It's this idea that a task requires an optimal amount of nervous system arousal. Too much and you get jittery, too little and you get lethargic. So it looks like an inverted U with the optimal performance at a middle level of arousal.
Although you're right it's more complicated. If you're an expert at a task, then you do better if you have increased arousal. This is because stress hormones cause new synaptic production to slow down and old networks/circuits to increase in their conduction/strength. So if you're an NBA player, and it's the playoffs, you're going to be better than normal - because you're an expert at the task.
If however the task is 'complex' (like, say, software development) then if you put pressure on people (or if they're putting pressure on themselves), the performance falls significantly. I forgot the name of the study, but this was shown cross-culturally. They gave poor Indian men (in India) 3-months salary to solve a decently straight forward lateral thinking puzzle. The men who got paid performed way worse, then the other men who weren't paid at all.
Sometimes that strikes me as oddly familiar. Paying someone to do something they love and putting too much pressure on them makes their performance degrade.
The highest pressure situation I was in was for fed gov approaching deadlines having missed milestones. The more pressure the put on, the calmer and more relaxed I became. Everything worked out fine in the end because the managers didn't understand investing in developing a framework. When that was done all the apps came together, more or less with outliers needing more exceptional additions.
Is it that simple? It certainly does not fit my experience. I type code for a living and have for 30 years so I'm clearly well rehearsed but I can't type well when being looked at.
This is psychology, there are always exceptions. Personally I don't have problems with the act of typing when being looked at, but I do have a tendency to get nervous to the point I'm having trouble speaking. Only happens with crowds I'm not familiar with though. But I know enough of sociolog/psychology to realize that doesn't mean any of the claims made here are false for the general public, or even for me in certain situations. I do have the impressions I'm sometimes better in sports when people are watching and/or cheering.
Perhaps that is not a performance, in the same sense as other stuff. e.g. do you get a bit of adrenaline kick in, are you motivated to impress them with your keyboard skills, etc?
For improving work you're good at - maybe not showing off, maybe people just avoid complacency/shortcuts, double check everything if they know their work will be checked. Like the value of code reviews that don't find any bugs.
