Classification of Data

Classification of data is the process of sorting a data set into categories based on what kind of information the values carry. The first and most important split is between qualitative (categorical) data and quantitative (numerical) data. Knowing which type you have decides which graphs you can draw and which statistics — like the mean or the mode — actually make sense to calculate.

Qualitative data describes qualities or categories — labels rather than amounts. It comes in two kinds:

• Nominal: categories with no natural order. Examples: eye colour, blood type, country of birth.
• Ordinal: categories that do have an order, but the gaps between them are not measurable. Examples: exam grades (A, B, C), survey ratings (poor, fair, good), t-shirt sizes (S, M, L).

You can count how many fall in each category, but you cannot meaningfully average them.

Quantitative data is made of numbers you can do arithmetic with. It also comes in two kinds:

• Discrete: counts that can only take separate, whole values. Examples: number of students in a class, number of cars sold.
• Continuous: measurements that can take any value in a range. Examples: height, weight, temperature, time.

The quick test: if the value answers "what kind?" or "which category?", it is qualitative. If it answers "how many?" or "how much?" with a number, it is quantitative. A postal code is numeric-looking but qualitative — it labels a place and you would never average it.

• Treating numeric labels as quantitative. Postal codes, jersey numbers, and phone numbers are qualitative — the digits are labels, not amounts.
• Confusing discrete and continuous. Counts are discrete; measurements are continuous.
• Averaging ordinal data. You can order it, but the gaps are not equal, so an "average grade" can mislead.

Once you know the data type, the next step is gathering it well — see data collection, the difference between a census and bias, and the main sampling methods.

Related lessons

• Statistics