Null hypothesis

• a null hypothesis is a formal statement that no difference exists between treatment groups (e.g. placebo versus treatment to be tested) or that there is no association between risk indicator and outcome variables. A null hypothesis is stating that any difference between the different treatment groups, or risk indicator and outcome variables, is due to chance alone.
• it is a hypothesis which is set up to be disproved by the results of a trial
• if the null hypothesis is true then the findings from the study are the result of chance or random factors. If a null hypothesis is rejected then, depending on the factors tested by the trial, there is evidence of a difference between treatment groups or an anssociation between risk indicator and outcome variables
• the predicted results given the accuracy of the null hypothesis are used to assess the statistical validity of the results of the trial

Relating the null hypothesis to type 1 error, type 2 error and the power of a study:

• null hypothesis - assumes that there is no difference between the groups (treatment versus placebo)
• type 1 error - alpha - the probability of rejecting a null hypothesis that should have been accepted, i.e. the probability of accepting an alternative hypothesis when an observation is due to chance
• type 2 error - beta - the probability that, despite accepting the null hypothesis (i.e. no evidence from the study results that there is a a difference between placebo and treatment), there really is a difference
• the power of a study is calculated as (1 - type 2 error)