Measuring and Describing Variables

Welcome

Topic:
How do we turn political ideas into things we can measure?

Example: Populism

Concept: Populism
Definition: Emphasis on “the people” vs. “the elites”

How we measure it: - Rhetoric in campaign speeches - Language in social media posts - Mentions of “corrupt elite” or “real Americans”

Multidimensional Concepts

Some concepts have multiple sides

• Democracy = contestation + inclusiveness (Dahl)
• Liberalism = economic + social dimensions

Each dimension needs its own definition and measure

Measurement Recap

• Conceptual definition: What we mean
• Operational definition: How we measure it
• Validity: Are we measuring the right thing?
• Reliability: Are we measuring it consistently?

What Is a Variable?

A variable is the measured form of a concept.

It has:

• A name

• Two or more values

• Usually numeric codes

Example

Concept: Political ideology
Variable name: Ideology
Values: 1 = very liberal, …, 7 = very conservative

We use variables to describe and analyze the political world

Levels of Measurement

Variables differ in how much information they contain:

1. Nominal
   
2. Ordinal
   
3. Interval

Nominal Variables

• Categories with no order
• Example: Party ID
  • Democrat = 1, Republican = 2, Independent = 3

You can say: “These are different.” That’s all.

Ordinal Variables

• Ordered categories, but spacing is unclear
  
• Example: Trust in government
  • Never / Sometimes / Always

You can rank responses, but not measure exact differences

Interval Variables

• Numeric values with meaningful spacing
  
• Example: % of votes, age, income

You can calculate exact differences: > 60% turnout is 10 points higher than 50%

One Concept, Multiple Variables

Political efficacy could be measured as:

• Yes/No → Nominal
  
• 1–5 scale → Ordinal
  
• Index of 10 items → Interval

More precision isn’t always better—depends on context

Describing Variables

Two big ideas:

1. Central tendency: What’s typical
   
2. Dispersion: How spread out the values are

Central Tendency

• Mode = most common value (all levels)
• Median = middle value (ordinal & interval)
• Mean = arithmetic average (interval only)

Dispersion

• Do values cluster or spread out?

Examples:

• Low dispersion: Everyone studies ~10 hours/week

• High dispersion: Some study 0, others 20+

Political Science Examples

Why It Matters

Your choice of variable type affects:

• What you can say
• What you can calculate
• How others interpret your findings

Thanks!