Statistics

Each post here takes one concept and answers the questions a thesis writer actually has about it. What it measures. How it is calculated, with a worked example small enough to check by hand. Which column it appears in when SPSS or SmartPLS prints it. And what counts as an acceptable value in social science research, which is not the same answer as in a laboratory. Every post ends with a paragraph on how to state that result in the results chapter, because knowing a coefficient is 0.68 and knowing how to write the sentence about it are two different skills.

A sensible reading order runs from describing data, to distribution and outliers, and only then to inference: p-values, hypothesis tests, and confidence intervals. If a reviewer has asked why you used 0.05 rather than 0.01, or why you reported a median instead of a mean, the answer is here rather than in a software manual.

Software procedure sits in the SPSS and SmartPLS categories so the posts here can stay on what the number means. Some overlap is deliberate: the reliability post here explains what internal consistency is, and the SPSS post shows you where to click.

Where to start. If you have just collected data, read the post on descriptive statistics so you can describe your sample with frequencies, means, and standard deviations. Then read the posts on normality and on outliers, because those two decide whether you are allowed to use parametric tests at all, and that decision belongs before any hypothesis test rather than after one comes out badly. Read the post on the p-value before you run anything, so that a small number does not get written up as a large finding. Finish with correlation and with sample size, the two concepts a committee is most likely to ask you to defend out loud.