Null Hypothesis

This is an empirically testable prediction relating to the population that is evaluated by considering a random sample from the population. Examples include, the proportion of ‘heads’ from a number of coin tosses, mean height, mean difference, etc. Typically, the researcher or analyst is seeking to reject, or refute, the null hypothesis, so that the more useful alternative hypothesis is accepted, for example, the proportion of heads is not equal to 0.5,  the mean is greater than 60, or there is a difference in population means. When experimental designs are statistically evaluated, the Null hypothesis is often referred to as the prediction of ‘no difference’. The exact form of the Null hypothesis will depend on the specific statistical test used, for example, the null hypothesis for the independent t-test would predict no difference in population means. The null hypothesis is sometimes written as H0: