Data types and measurement scales

Learning objectives

• Classify data as continuous, count, binary, nominal or ordinal.
• Match data type to appropriate chart and test.
• Avoid discarding information by premature coding.

Why scale matters

Continuous data (cm, seconds, grams) carry the most information. Count data record how many. Binary is pass/fail. Nominal has no order; ordinal has order without equal spacing.

Converting a chest-width measurement to pass/fail before analysis throws away the signal needed to see drift and centring.

Garment-factory example

Chest width in cm is continuous; defects per garment is a count; ‘shipped on time yes/no’ is binary; defect type is nominal; shade grade A/B/C is ordinal.

Method

1. State the analytical question first.
2. Identify the unit and the scale.
3. Keep continuous readings when the instrument allows.
4. Document any coding rule so it can be reversed later.

Common mistakes

• Recording a caliper reading as pass/fail only.
• Averaging category codes as if they were numbers.