Thống kê và dữ liệu
Moderation asks when an effect is stronger or weaker; mediation asks through which pathway it works. Interactions, simple slopes, bootstrapped indirect effects
Thống kê và dữ liệu
Bayesian statistics answers the question many people think a p-value answers: given these data, how believable is the hypothesis? Priors, posteriors, credible i
Thống kê và dữ liệu
The ordinal-versus-interval argument about Likert data has run for decades. The practical answer lies in separating a single item from a multi-item scale, and i
Thống kê và dữ liệu
When your question is “how long until it happens”, such as time to PhD completion or to dropout, the average time is the wrong number. Censoring, Kaplan–Meier c
Thống kê và dữ liệu
Surveying 800 pupils in 20 classes does not give you 800 independent observations. Ignoring nesting makes standard errors falsely small. ICC, the design effect,
Thống kê và dữ liệu
The chi-square test is the most-run test in survey theses and among the most misread. Expected counts, when to use Fisher, reading residuals to find which cells
Thống kê và dữ liệu
Shapiro–Wilk gives p < 0.05 and many researchers jump to Mann–Whitney. But normality tests usually answer the wrong question, and Mann–Whitney does not compare
Thống kê và dữ liệu
A meta-analysis turns 20 small studies into one estimate, or pools apples with oranges and reports a very confident number. Weights, fixed versus random effects
Thống kê và dữ liệu
When the outcome is pass or fail, dropped out or stayed, linear regression no longer fits. How to read logistic coefficients, convert them to odds ratios, avoid
Thống kê và dữ liệu
What does a coefficient of 1.8 mean when five other variables are in the model? Reading each coefficient, reference groups, R squared, residual checks and multi
