The analysis itself
Statistics
Every test a postgraduate thesis normally needs, with the checks that decide whether the test was the right one already run.
You choose what you are trying to find out. ZStat picks the procedure, runs the assumption checks whether or not you asked, and writes the result as a sentence you can paste into a chapter with the numbers already formatted to APA 7.
What you can do
Two groups or twenty, related or independent, normal or badly skewed. t tests, ANOVA in its factorial and repeated-measures forms, and the non-parametric equivalents when the assumptions do not hold.
Correlation, partial correlation, linear and logistic regression, ordinal and multinomial outcomes, mediation and moderation.
Reliability, exploratory and confirmatory factor analysis, structural equation models, item response theory, and measurement invariance for when you need to compare groups on a questionnaire.
Survival analysis, time series and forecasting, and panel data with the Hausman test choosing between fixed and random effects rather than you asserting one.
Missing data patterns, weighted survey estimates, bootstrapping, exact tests for small samples, and meta-analysis across studies.
What it refuses
- A t test on three groups. It names the groups it found and points at the one-way ANOVA.
- A correlation reported without its reliability, when a noisy scale caps the correlation you could possibly find.
- Reporting a non-significant difference as evidence that two things are the same. That is a different test with a margin you have to state.