≡ Menu

Standard Cautions That Apply To Studies Reported On PsyBlog

Standard Cautions That Apply To Studies Reported On PsyBlog post image

Correlation does not equal causation and more standard disclaimers.

Below I’ve listed some ‘standard disclaimers’ that should be applied to some of the studies I mention.

Occasionally I mention these in passing, when they are relevant, more often I don’t.

Most of these disclaimers would be introduced to first year psychology students early on in the course.

I don’t like to repeat them every time as it gets repetitive, so here they are in one place.

  1. Correlation does not equal causation. If a study has a correlational design, you can’t conclude that one thing causes the other. If you see the word survey then it’s probably a correlational study.
  2. One swallow does not a summer make. One study can’t prove a theory true or, for that matter, prove a theory wrong.
  3. People are different (I). Psychologists are generally interested in how everybody’s minds work, on average. Individuals can be as unique and different as two snowflakes.
  4. People are different (II). Lots of studies are done on young, white, middle-class college students. Lots of real-live people are not white, middle-class college students or even all that young. Does that mean the results of the studies should be binned? Not necessarily, but it is something to bear in mind.
  5. It’s the effect size, stupid. Just because a study finds a statistical difference, it doesn’t mean that it makes a real-world difference. Statisticians use something called an ‘effect size’ to quantify this. I almost never mention these.
  6. They only measured two things. In the simplest studies, scientists measure two things (let’s say happiness and long life), then, after a survey, pronounce a relationship between them. Yes, there could be all kinds of other things going on that weren’t measured, it’s true.