Types of variable, and why the type decides the test
Numerical or categorical; and if categorical, ordered or not.
Numerical (quantitative) — a measurement. *Continuous* can take any value in a range (height, HbA1c, VAS score). *Discrete* counts whole things (number of admissions).
Categorical (qualitative) — a label. *Nominal* has no order (blood group, district). *Ordinal* has an order but unequal gaps (mild / moderate / severe; a five-point Likert item).
Why it matters
This is the first question every test-selection rule asks. Two groups with a numerical outcome points at a t-test; two groups with a categorical outcome points at chi-square. Get the type wrong and every downstream choice is wrong with it.
An example
In a knee osteoarthritis study:
• VAS pain 0–10 — numerical, continuous
• Range of motion in degrees — numerical, continuous
• Number of flare-ups in six months — numerical, discrete
• Kellgren–Lawrence grade I–IV — categorical, ordinal
• Affected side (left / right / both) — categorical, nominal
• Improved (yes / no) — categorical, nominal
Common mistakes
- Averaging an ordinal scale as though the gaps were equal. The distance from 'mild' to 'moderate' is not the distance from 'moderate' to 'severe'.
- Cutting a numerical variable into groups without a reason. It throws away information and usually costs you power.
- Coding categories as 1, 2, 3 and then treating them as numbers.
Read next
- How the right statistical test is chosen — Four questions decide it, and none of them require you to know statistics first.
- Normality, and what to do when it fails — What the assumption actually means, how it is checked, and why failing it is not a disaster.
Chavery Research Companion applies this to your own study: it asks the questions in plain language, checks the assumptions against your data, recommends the test, and writes the sentence that reports it. Start free — planning and the master chart cost nothing.