Demonslayer
Well-Known Member
To control for it you would, quite literally, have to comprise your samples of 100 and 100 of people tho had the same "sorts" of parents. "N" number of people in each sample whose parents who provided mediocre or sub-standard "love" for instance. But how exactly does one go about gauging something like that? Well... I don't think you can.
But what you're saying here is that no behavioral studies are possible because of variables. Of course this isn't true, valid studies can be done controlling as many variables as possible and using statistical significance to answer for the rest.
Statistical significance is a threshold of difference between the results of two groups. If there is a difference between two groups in a study that is under the significant threshold, the idea is that the difference could have been caused by other variables. If you're over the statistically significant threshold, the difference is considered too great to be caused just by other variables, and the tested variable is very likely at least in part the cause of the difference.
I mean we know smoking causes cancer. There are, of course, may other factors in whether or not a person develops cancer. Where you live, what your job is, how much you drink, your diet, family history, etc. There are lots of variables as you describe in the spanking study. But this doesn't mean we can't scientifically link smoking to cancer, we certainly can. The reason is when we compare smokers to non-smokers, the results are statistically significant which means the other factors could not have been the only reason for the results and that the tested factor (smoking) is most definitely in play.
That doesn't mean we won't see a smoker who never gets cancer or a non-smoker who does.
Still interested in the "why" questions I asked you. There has been a meta-study done over 50 years showing a link between spanking a statistically significant increase in 13 negative behaviors. You'd bet money against the study.
Why?