Statistics guides

A practical decision guide: from the shape of your data to the right test, in four questions.

Two groups, different subjects: run it, check assumptions, report it — with real data and real numbers.

Before and after, left and right, two instruments on the same samples: when each subject supplies both values, pair them.

Four fertilisers, one question: which ones actually differ? From raw data to a lettered, journal-ready figure.

Two factors at once: main effects, the interaction, and the two-tier letters that summarise it all.

Three factors varied together: three main effects, three two-way interactions and the three-way interaction, in one run.

The same subjects measured over time: why ordinary ANOVA is wrong for this, and what to run instead.

Every subject measured under every combination of two factors — the design, its own error terms, and the sphericity check.

One between-subjects factor (treatment group) and one within (time): the split-plot design most longitudinal experiments are.

What the letters above your bars mean, how they are computed, and how to put them on your own figure in a minute.

Fitting a line, reading the slope, and knowing when a line is the right model at all.

A dose series with a plateau at each end: fit the four-parameter logistic and read the EC50 with its confidence interval.

A growth curve, a release profile, a temperature log: choose the window and get the slope, its error and its significance.

When the outcome is a category (germinated or not, passed or failed), model the odds of it from one or more predictors.

Several variables measured on the same samples: every pairwise correlation, its p-value, and what to make of the pattern.

Two multivariate tools that look alike and answer different questions — with a worked PLS example.

What the colours, the two trees and the z-scores mean, when to use it beside PCA, and how to read the order table.

Entering data, the templates, labels, preview and confirmation — the workbench before any statistics.

Every option of the Figure styling panel, tab by tab: text, axes, colours, marks, letters, reference lines, panels, legend, frame and size.

Collecting figures into a multi-panel layout: adding, arranging, numbering, images in panels, panel styling, removing.

Figure downloads and their formats, the full Word report, and where the Methods text and the citation sentence appear.

The click path for PCA, PLS, the correlation matrix and the heatmap: the data shape in, the options, and what each screen produces.