A hypothesis test is a procedure that decides whether the observed data is compatible with a null hypothesis (H0), generally the absence of an effect. It is set against an alternative hypothesis (H1). You set a risk threshold alpha, then compute a test statistic and a p-value. Rejecting H0 does not prove H1.
Example
H0: the mean is 50; H1: the mean differs from 50; threshold alpha = 0.05.