Can love and friendship replace money?

[physicsdude30]

So if the only basis you have for understanding social-desirability bias is what you've read about it in books and articles, then you should really do the effort of self-reflecting and identifying with it in practice.

One of my past jobs was telephone surveys for a market research company. From personal experience I can say that the more "distance and anonymous they feel from you", the less likely they'll feel to impress you with their answers or social desirability bias. If you're in person face to face in a case study interview, they're really going to be feeling the social desirability bias. As far as being no nonsense down to earth, you can even see with your own ideas that the more anonymous/distance they feel, the less likely they feel to give an answer because they feel to impress.

You say that is just my personal experience and interpretation? Taking that logic that's why I'm saying relying on case studies alone may not necessarily give the big picture. Different researchers looking at case studies are going to see the exact same situation and interpret it much differently, in addition to plain having different situations. That's where looking at correlational studies has a big big advantage. To demonstrate this principle, would you agree smoking is bad for you? Along these lines, I talk to one person and they say they know someone who always smoked and developed lung cancer and died early. Then I come across another who says he had a grandpa who lived to be 100 and smoked most days of his life. He says that he thinks the Medical Field is lying when they say smoking is bad for you, and that you have to think for yourself rather than what the Medical Field says. What would you think if you heard on the news someone saying smoking is great for you because of a case study where someone lived to be 110 and smoked their whole life? That's what happens when relying on case studies alone, different experiences and interpretations. However, looking at correlation gives us a "bigger picture" that smoking is strongly related to premature death, lung cancer, certain heart diseases, etc. Although you can't prove for sure without using an experimental-control study (unethical to perform with smoking), when controlling for many third variables the correlation is still there.

That's why I'd be interested in setting up a study where you can look for correlations and control for third variables to see if there's any left over, in order to see a bigger picture. Also, although correlation doesn't prove causation, it's a requirement, and so if there's a correlation after controlling for many other variables it would give us more confidence (or the opposite if no correlation). Although you can't prove for sure, testing something with the Scientific Method is better than nothing at all.