A laboratory manual for social psychology.Experiment 3: Polling—Reliability of sampling results.
Abstract
In most investigations in the field of psychology, as, indeed, in most other fields, the question arises toward the end of the work as to how much confidence can be placed in the results. Assuming that the experimenter knows what he is doing, that his experimental design is adequate, and so on, there still remains the question as to whether his results are due to chance or to some difference between his experimental groups. Such questions are answered by statistical procedures. But even when the statistics indicate a real difference, the experimenter has to decide what caused the difference. (PsycINFO Database Record (c) 2009 APA, all rights reserved)