Asking about salaries (like asking about anything else) produces unreliable information; i.e., information that may be, and often is, inaccurate, incomplete, and/or biased. I make this point in my book, The Problem with Survey Research, and so does Bob Weinstein, pointing out that we shouldn’t assume surveys are accurate, but rather that “most surveys are awful”. Many survey researchers, themselves, admit that their asking efforts are “not scientific”. Weinstein also contends, and correctly so, that Internet surveys are essentially “worthless”.
Rather than asking about salaries, or anything else, we should use content/document analysis, observation, and other “proper” –as I call them in Part 6 of my book–methods of data collection and proper research designs. These are the procedures that produce reliable information.