Directional hypothesis

A directional hypothesis is an alternative hypothesis where the researcher predicts the directional nature of the predicted outcome before data collection. For example, employing a correlation where the correlation was predicted to be positive or negative, one or the other, but not both. Similarly, predicting a mean difference where a specific  mean is predicted to be higher than than the other. Normally, a directional hypothesis would require a priori justification rather than being arbitrarily constructed to suit the data. The use of a directional, as opposed to a non-directional hypothesis, will lead to a different calculated significance level (p-value).