significance test


Also found in: Dictionary, Medical, Wikipedia.
Related to significance test: significance level, T test

significance test

(STATISTICS) a test designed to assess whether an observed (numerical) result can have occurred by chance. The result of the test is expressed as a statistic (e.g. t-ratio, F-ratio) which can be assessed against different levels of probability. It is usual to accept a level of probability of 0.05, i.e. that there is only a 5% probability of the result having occurred by chance.

Examples of significance tests are the t-test (parametric) and the Wilcoxon (non- parametric) (see STATISTICS AND STATISTICAL ANALYSIS). These tests are designed to test for the significance of the observed difference between two groups of data. For example, in social survey work two samples may be taken, racial attitudes in cities with and without ethnic minority problems, perhaps. There may be an apparent difference (numerical) between these groups, but the groups were samples and therefore the data is subject to SAMPLING ERROR. The difference between them must therefore be tested to see if there is a statistically significant difference between them. Significance tests are designed to set up a NULL HYPOTHESIS, stating ‘no difference’, and the test result either confirms or disconfirms this.