Likert Scales

Summated rating statements

A Likert scale measures an attitude or trait by asking respondents to rate a series of statements on an ordered response scale, typically ranging from strongly disagree to strongly agree; scores across items are then summed or averaged. Its simplicity and versatility have made it one of the most widely used measurement tools in the social sciences. A long-standing methodological debate concerns whether the resulting data are ordinal or can legitimately be treated as interval for parametric analysis.

Concept Definition

A Likert scale is a measurement approach that presents respondents with a set of statements (items) related to the construct of interest and asks them to respond to each using an ordered response format. Response options are graduated, typically consisting of five or seven categories, and can reflect direction (strongly disagree to strongly agree) or frequency (never to very often). Using multiple items together yields richer measurement and higher reliability than a single question could provide. The total or average score across the scale represents an underlying latent variable that is not directly observable.

How It Works: Structure and Types

Designing a Likert scale begins with clearly defining the construct to be measured and writing items that reflect it. All items use the same number of response categories, which are assumed to represent something close to an underlying continuum. Items may be worded positively or negatively; negatively worded items are reverse-scored before analysis. The five-point format is most common, but four-point formats (forcing a position by removing the neutral midpoint) and seven-point formats (capturing finer distinctions) are also widely used. Scale validity and reliability are evaluated through item-total correlations, internal consistency coefficients such as Cronbach's alpha, and confirmatory factor analysis.

A Concrete Example

Consider a researcher measuring students' attitudes toward statistics. Items such as 'Statistics is useful in my daily life', 'I enjoy solving statistics problems', and 'Thinking about taking a statistics exam makes me anxious' could be presented on a five-point scale. The third item is negatively worded and therefore reverse-scored before analysis. Each student's scores across all items are summed or averaged; a high total score indicates a favorable attitude. This structure allows the researcher to represent the latent attitude construct far more reliably than any single question could, capturing multiple facets of the same underlying construct.

Common Pitfalls and Best Practice

The most frequent misconception is treating a single Likert item as a scale; by definition, a scale comprises multiple items summed together. The measurement level debate is also critical: because responses are technically ordinal, the use of means and parametric statistics is contested, though widely accepted in practice when items are numerous and distributions approach normality. The number of response options, the presence or absence of a neutral midpoint, the number of items, and the clarity of item wording all directly affect data quality. The midpoint is also ambiguous: some respondents use it to signal uncertainty while others use it to signal indifference.

Key terms

Item: A single statement in a scale; multiple items together constitute the full scale.
Summated Score: The sum of scores across all items; the core output of a Likert scale.
Reverse Scoring: Flipping scores on negatively worded items before analysis to ensure directional consistency.
Internal Consistency: The degree to which scale items measure the same construct, assessed with Cronbach's alpha.
Ordinal Measurement: Measurement level where categories are ordered but intervals between them cannot be assumed equal.