LESSON 41 - LIKERT SCALE || WHAT IS LIKERT? || DOES IT COLLECT CONTINUOUS OR CATEGORICAL DATA?

Likert scales turn agreement with statements into structured responses, but whether those responses should be treated as categorical or continuous data determines how researchers analyze them. The key decision happens during questionnaire design: if Likert items are used as separate response categories, the data are handled as ordinal (typically with modes and frequencies). If multiple Likert items are combined to represent a single variable, the resulting scale can be treated as interval data, allowing parametric statistics such as means and standard deviations.

Each Likert item commonly uses a five-point format that captures positive or negative attitudes toward an object. For positively worded statements, “strongly agree” is scored 5, “agree” 4, “neutral/undecided” 3, “disagree” 2, and “strongly disagree” 1. For negatively worded statements, the scoring direction is reversed so that higher scores consistently reflect the same underlying direction of attitude. The lesson stresses that these numeric scores do not automatically make the data continuous; they only carry quantitative meaning when the scale is constructed and analyzed as interval-level measurement.

Constructing effective Likert items requires clear, respondent-friendly wording tied directly to the research problem. Statements should be short and simple so respondents can answer without fatigue. Factual statements should be avoided because they don’t require judgment—respondents can simply agree. Double negatives should be eliminated because they force confusion about what is being endorsed. The scale should balance negatively oriented and positively oriented items; negative items must be reverse-coded so the overall scale remains coherent.

Item quantity matters, especially when aiming for inferential analysis. The scale should include at least 10 statements, with a practical upper range around 20–25, and the items should not be overly long. Researchers should also avoid irrelevant statements that don’t map to the research questions, and avoid items that almost no one will endorse—often a sign of abstract wording or unclear language that leads to skipped responses. Double-barreled questions are also discouraged because they test two variables at once, which can blur interpretation.

Finally, the lesson tackles three common myths. First, Likert items are not independent; they are linked by the underlying conceptual and empirical structure of the variable they measure, so individual statements can’t be treated as autonomous. Second, Likert items shouldn’t be analyzed in isolation; they belong to the questionnaire and should be analyzed together as a set addressing a variable. Third, building a Likert scale requires more than converting variables into questions—researchers must revisit variables, indicators, and research questions to craft items that truly measure what the study intends to measure. The session ends by previewing interviews as the next data collection method.