OrientationObservationIn-depth interviewsDocument analysis and semiologyConversation and discourse analysisSecondary Data
SurveysExperimentsEthicsResearch outcomes
Conclusion
8.4.3.6 Significance tests This Section explains and illustrates the different forms of significance tests (of null hypotheses).
The principles of hypothesis testing have been explored in previous Sections, which involved using tests to see whether a claimed population mean is the same as the true population mean on the basis of the sample taken at random from the population. There are many more tests for different situations.
See here for a chart showing which test is appropriate for different scales of data and sample size.
Section 8.4.3.6.1 examines parametric tests, which test parameters of samples, such as mean, proportion, standard deviation, correlation coefficient.
Section 8.4.3.6.2 examines some non-parametric tests, which test the overall distribution (shape) of a sample against another.
It is important to be aware that a statistically significant difference is not the same as a substantive difference. A significant difference is based on probability and indicates whether a sample is likley to come from a specified population or whether two or more samples are likely to be different. This is a different concept to substantive difference, which is about considering whether a difference is substantial or not, irrespective of where it is statistically different. See Statistical vs. Substantive Significance (accessed 30 April 2020) for a further brief explanation (also available here).
