Statistical significance: Significance that persists even after eliminating ‘chance’.
Level of significance: An indicator/ measure of the magnitude of statistical significance, this is also called ‘alpha’. Typical values range from 5% (which corresponds to a p value of 0.05) to 0.01% (p value of 0.0001).
Test of statistical significance: A test that helps compare one’s finding(s) with some other value(s) in order to determine if statistical significance exists at a given level of significance.
Parametric tests (of significance): Tests that assume the values are normally distributed. Example: student’s t test
Non-parametric tests (of significance): Tests that do not require/ assume the data to be normally distributed. Example: Mann-Whitney U test.
Hypothesis testing: A way of checking if a given statement may be true.
Null Hypothesis: A statement that says there is no difference between two alternatives.
Alternative Hypothesis: A statement that says there is a difference between the alternatives.
Two-sided hypothesis: A statement that does not attempt to predict which alternative is better or worse; merely says the two alternatives are not equal/ same.
One-sided hypothesis: A statement that clearly says one alternative is better, indicating which way the results might be expected to lean.
Null Hypothesis Significance Testing: A specific way of testing a statement. It involves stating the Null Hypothesis
p value: The probability of obtaining similar or more extreme results if the null hypothesis is true.
Clinical significance: Distinct from statistical significance, this indicates the clinical (real-world) importance of a finding. Statistically significant findings may be clinically irrelevant, and vice-versa.