There are two types of hypotheses. The most basic is a data level hypothesis predicting how the data will come out (if the hypothesis was made before the data is collected) or reported as a confirmed or disconfirmed hypothesis after the data is collected. Those are in the first sentence of these tutorial problems, the "facts" part of the statements, which is what you want to choose as the hypothesis. The other type of hypothesis is a theoretical hypothesis, an explanation of the "why" of the  results (see the discussion below on Osborn, 1996.) When we consider alternative explanations in correlational studies we call them Plausible Rival Alternative Hypotheses -PRAHs when we come up with an educated guess different from the experimenter's explaining why the data came out the way it did. PRAHs exist because compared to true experiments, correlational studies have missing information that make them inherently ambiguous.  Strictly speaking there are no PRAHs in true experiments so as soon as you determine your problem is a true experiment you can quit worrying about PRAHs

However, in both quasi experiments and experiments there may be some ambiguity about which out of a variety of theories explains the facts of the data.  So data that appears to confirm a theory may have other theories that might explain it.  Therefore, the great philosopher of science Karl Popper pointed out the real test of a theory is when its prediction could be contradicted (or disconfirmed as we say) and it is not disconfirmed. So a theory is stronger each time it survives a possible failure. Here is a concrete example of this process.  When Osborn's 1996 experiment found that makeup increased a woman's attractiveness he claimed this data showed beauty is a socio-cultural attribution, a personal identity explicitly chosen by the woman when she puts makeup on and not a biological factor. The biological beauty camp disputed his interpretation of the results of his study and said makeup works to increase attractiveness by simulating a biologically healthy appearance, covering up wrinkles and other imperfections. What would it take to choose between these theories of beauty, the socio-cultural or biological?  We need a study that had the same women rated with (A) makeup that stealthily concealed flaws but looked like no makeup and also rated (by another group of raters) with (B) makeup that was obvious. If the biological beauty group is right group A would be rated more highly than group B thus disconfirming Osborn's socio-cultural theory of beauty.

Return to Plausible Rival Alternative Hypotheses       Return to Home Page