[QUOTE]Originally written by gsociology on 2003-12-08 11:18 PM Hi The usual description is something like the following: Suppose you made a survey asking people what they thought of G.W. Bush, approve, neutral, disapprove. You got the same answers regardless of how you word the question, which surveyor asks, which day, and so on. This is all about reliability. Your survey is reliable. Here is what differentiates validity. Suppose you were a republican, and in your mind, GW Bush = the republican party. So to your thinking, you are asking people what they think of the republican party. This, however, is not really true. GW Bush does not represent the repub party. So, your survey IS a valid measure of the qeuestion you are asking: what people think of GW Bush, but it is NOT a valid measure of what you really want to ask: what people think of the republican party. Hope that helps. Also see statnotes on reliability http://www2.chass.ncsu.edu/garson/pa765/reliab.htm validity http://www2.chass.ncsu.edu/garson/pa765/validity.htm susan losh's lecture notes on reliability and validity http://edf5481-01.fa02.fsu.edu/Guide3.html gene shackman Free resources in social research methods http://gsociology.icaap.org/methods/ -- The global social change research project http://gsociology.icaap.org/[/QUOTE]