What is Interval Scale?
An interval scale is a type of measurement where values are ordered and the gaps between values are equal, but there is no meaningful zero point. It lets you compare differences (add/subtract) but not form true ratios (you can say one score is 5 points higher, not 'twice as much').
Interval scales assign numbers to responses so that the order is meaningful and the distance between adjacent values is consistent. Common examples are temperature in Celsius or dates on a calendar. Because an interval scale lacks a true zero, multiplying or dividing values (saying one value is twice another) is not meaningful. In survey work, a 1–5 satisfaction scale is often treated as interval for calculating averages, though strictly speaking it is ordinal unless validated as having equal intervals.
Usage example
A school runs a parent satisfaction question with answers 1 (Very dissatisfied) to 5 (Very satisfied). Treating that as an interval scale lets the admin report the average satisfaction score (e.g., mean = 3.8) and compare differences between groups.
Practical application
Knowing a measure is interval guides which statistics and tests are appropriate. You can compute means, standard deviations, t-tests or ANOVA on interval data (with reasonable assumptions), but you should not report ratios or claim one score is twice
another. For Hearo users this matters when designing response options, comparing group averages across languages, or deciding whether to aggregate open-text sentiment into numeric scales — and when explaining results to stakeholders so conclusions stay accurate and defensible.
FAQ
Is a Likert scale an interval scale?
Strictly speaking, Likert items are ordinal because they show order but not guaranteed equal spacing. In practice many practitioners treat 5- or 7-point Likert scales as interval when the categories are symmetric and well understood, which allows calculation of averages and common parametric tests — but you should be cautious and consider non-parametric alternatives if in doubt.
Can I calculate an average on interval data?
Yes. Calculating means and standard deviations is appropriate for interval data and is commonly done to summarize responses. Just avoid interpreting ratios (e.g., twice as much
) because interval scales lack a true zero point.
How is interval different from ratio scale?
Both have ordered values and equal intervals, but a ratio scale has a meaningful zero (e.g., weight, height). With ratio data you can meaningfully say one value is twice another; with interval data you cannot. This difference affects what mathematical operations and interpretations are valid.
Why does this matter for multilingual surveys?
When you translate a survey, preserving equal perceived spacing between response options matters if you plan to treat the answers as interval. Poor translation or inconsistent phrasing across languages can shift how respondents interpret categories and break the equal-interval assumption, which can bias comparisons of means between language groups. Hearo's translation and review workflow helps maintain consistent wording so numeric comparisons remain meaningful.