It's not going to tell us much because we don't know the standard of code quality that respondents hold themselves to.

True, but easier to see consensus than reading through a thread that is filled with a variety of posts in terms of usefulness.