Empirical and Normative Inquiry: Concepts, Evidence and Comparison
8 min read
Lesson objectives
- Distinguish empirical, normative and mixed claims.
- Build a simple operational definition of a political concept and state what it does and does not measure.
- Explain the limits of generalization, association and causation in a simplified political example.
Before disagreeing about the answer, what kind of question is it?
Political inquiry distinguishes empirical questions that can be examined with evidence from normative questions about what ought to be. The two can interact, but evidence should not be replaced by a value judgment, and a moral value should not be presented as though it were a statistical finding. This distinction is foundational to neutral reading of political evidence.
From concept to measure: operational definition
Terms such as “participation,” “trust” and “polarization” are broad. Before measuring them, we need an operational definition specifying what counts as evidence of the concept. “Participation in a local session” could be measured by attendance, the share making a contribution, or the number voting. Each measure captures a different aspect, so state what the indicator does and does not measure.
Description, association and causation are not the same claim
Suppose students who attended an orientation workshop voted more often than those who did not. That is an association in the data. It does not by itself show that the workshop caused participation; people already interested in politics might be more likely both to attend the workshop and to vote. Causal analysis needs a stronger design or additional evidence that reasonably addresses alternative explanations.
Generalization needs stated boundaries
A generalization summarizes a pattern beyond individual cases, but it does not mean every person matches an average or that a relationship holds in every context. State the population studied, time period, selection method, variable definition and major sources of uncertainty. Data from one organization should not automatically become a universal law.
| Type | Example | What is needed? |
|---|---|---|
| Empirical descriptive | How many people participated? | Data and a clear measurement definition |
| Empirical explanatory | Which factors are associated with participation? | Comparison, evidence and alternative explanations |
| Causal | Did an intervention change participation? | A design supporting a reasonable causal comparison |
| Normative | What level of participation is fair or desirable? | An explicit value principle or argument |
| Bounded inference | What can current evidence support? | Explicit scope, limits and uncertainty |
Lesson summary
An empirical question seeks a description or explanation that can be examined with evidence, while a normative question asks what ought to be. Moving from a broad concept to an indicator requires an operational definition; association is not causation; and generalization requires explicit boundaries around population, time and measurement.