Whenever you say statistics one of two things happens, people either assume you work for a politician and everything you are about to say is lies, which is not what I do so I will say no more there, or they think you are about to drone on about boring mathematical principles and nothing will make sense. Give me the finding, show the brain scans that light up and tell me that we have cured every brain disease ever, even though we haven't and you probably already know that before the new alert has finished.
There is no way around it; unfortunately, psychology and neuroscience are statistics, they wouldn't exist in their current form without them, and there is no way I can teach you to navigate all the findings, news reports, and phony business ideas without explaining at least a little bit of statistics. But according to good teaching etiquette, I am going to explain why this matters. Well, imagine we want to test memory, we want to see if a new drug makes people remember things better. We could give you the drug, but the placebo effect gets in the way. For some reason, and no one really knows why humans respond to fake treatments as they expect them to work, our imagination goes wild, and just believing a treatment works is enough.
For those that haven't heard much about the placebo effect I am going to heavily stress here it has its limits. The placebo effect can only mimic weak effects; it can make you think a cheap painkiller is taking effect when really it is sugar, or make you think you have better memory and maybe remember one more thing. The placebo effect isn't strong enough to cure cancer or provide a perfect memory or alter someone's personality forever. This is obvious when you think about it because medicine exists, if all that a doctor had to do to cure cancer was say this pill cures cancer, we wouldn't have a disease. But we do.
Ok, so we need to provide a fake pill and a real pill and see how our two groups respond. That is sensible and no stats so far. We also probably need a bunch of people; we can't just get one person to remember some words and randomly grab another person. What if one just has a really good memory or needs to use their memory for their job so is more trained. We would end up testing the person's, what's called individual differences, and not the pill. And even if we got two identical people, what if they both have a bad day, or aren't very good at this test, etc. It is much better to go get hundreds of people and ask them to take our real and fake pills and do a memory test.
Now we have 100 people who have their scores for the real pill and one hundred who took the fake pill, but neither group know what pill they took because that would ruin the reason for a fake pill. You can tell people after, but the deception is allowed by ethics during the test. Ok so each person did 20 memory tests, and we have 200 participants, so we now have 4000 memory tests to go over. That is too much. We should take the average; see how they did in general. So we take the average by adding up someone's scores and dividing by the number of tests they did, the same idea as figuring out your average electric bill by adding them all up, divide by twelve, and that on average is what you spent per month. Now we have 200 results, 100 averages for the real pill and one hundred averages for the fake pill.
We could just do another average, but that doesn't say much. Here is why. People vary, and sure if the real pill people could remember 10 items on average and the fake pill people remember 5 items, then we can be pretty sure the real pill improved memory. That isn't how it usually works; it may be that the fake pill made people remember 9.2 items on average and the real pill 9.5 items. Effects in psychology are subtle; memory is a complex system with lots of moving parts in the brain, we only gave one drug so we likely only effected a couple of parts of the brain and had a small effect on memory overall.
But the real pill score is higher right? Maybe not, here comes the stats. Like I said people vary, the results get scattered all over a graph, and maybe that .3 more items remembered isn't a real difference, it is just bad luck our groups weren't perfectly separated. We don't know until we test, even the test to know has to go through statistics so you can't escape this problem. We need to do tests that let us know how likely it is these two groups are different, give me a probability. It will never be perfect, but at least I know, in psychology as a field we chose that a test can be wrong 5% of the time, in medicine, it is 1% because the stakes are higher and other fields pick their number. There is no answer written in stone; you just have to live with it.
We give a stats program the numbers, and it says yes, this could be one of that 5% of times I'm wrong, but these two groups do have a difference and since the only difference is the pill, the pill probably caused it. That is what stats shows, that is the reality of the matter. It is important you know this, I spent a lesson talking about this because this is at the heart of all psychology and neuroscience. The science still works, but like the machines, it has limits.
Statistics is something you spend over a quarter of your time on in a psychology degree, and more in a neuroscience degree because this matters so much. If you want to know more, I promise it really is fascinating.
