Numerical variables are measured or counted (age, weight, number of children). Categorical variables put people into groups (sex, blood group, district).
Why it matters
The type decides which summary, chart and test you can use. A mean of blood groups is meaningless.
Key ideas
- Numerical: continuous (height) or discrete (count)
- Categorical: nominal (no order) or ordinal (ordered)
- Codes like 1/2/3 are still categories if they only label groups
- Summarise numbers with mean/SD or median/IQR
- Summarise categories with counts and percentages
Real-life research example
"Number of antenatal visits" is numerical (discrete). "Education level: primary, secondary, higher" is categorical (ordinal).
When to use
- Decide the type of every variable before analysis
When NOT to use
- Don't average codes of categories
- Don't treat a numeric ID as a measurement
Quick takeaway
Numbers get means; categories get counts and percentages.
Try it yourself
Paste one column of numbers and one column of words, one at a time, and check how each is detected and summarised.
Model answer
Numbers are summarised with centre and spread; words are summarised with counts and percentages. A numeric-looking code (1 = male, 2 = female) is still categorical.