Statistical Research and Ethics

Question

Why is it important that statistical research follow ethical guidelines? What are some potential consequences of not following ethical statistical practice at work?

Answer 1

Hypothesis testing is misleading if you deliberately choose a null and alternative hypothesis that will give you the result you want, before looking at the data or running the test

Statistical graphs can be misleading if you use different units on the axes. You can make a fairly small change look huge by altering the units

As Mark Twain said, there's "lies, damned lies and statistics"

Cigarette companies used statisticians to make it look as if cigarettes did not cause cancer, when they knew it did. Statistical models depend on underlying assumptions. You can get most any result you want by altering the assumptions, some not explicit

Some blood pressure statistics seem unethical, as if they are really taking measurements without the BP meds, you're endangering lives, risking heart attacks. Yet that's how some measurements appear, unless they just made up the numbers, which is another unethical behavior. Mendel with his genetic experiments with plants had numbers that looked too perfect. He was right in the results, but he seems to have faked the numbers to fit what believed to be true.