Descriptive and Causal Questions

The text below is a transcript of the video above, edited with AI assistance for clarity and flow. Where the two differ, the video is the authoritative version.

You’ll learn to

Empirical research questions can be descriptive or causal. In this video you will learn to:

  • Distinguish descriptive and causal questions.
  • Explain what kind of claim each type of question supports.

Why care about this distinction?

Why should we care about distinguishing between causal and descriptive questions? Descriptive questions ask what the world looks like, whereas causal questions ask what would happen if we intervened.

Descriptive questions

Descriptive questions can be about distributions, trends, or correlations.

A descriptive question about a distribution would be: what percentage of countries are democratic? In other words, what is the distribution of regime types across the world?

A question about a trend, and a pressing question of today, would be: is there a global trend towards democratic backsliding? Are we observing this change over time? Can we describe a change across time, with more countries experiencing democratic backsliding than before?

A descriptive question about a correlation would be: is democratic backsliding more prevalent in countries with higher political polarisation? Describing a correlation does not mean that we are talking about causation. In a question like this, we are not necessarily asking whether political polarisation causes democratic backsliding, or whether democratic backsliding causes political polarisation. It is simply a descriptive question about a correlation between democratic backsliding and political polarisation.

These questions are foundational for scientific research. They are incredibly important. If we want to understand the state of democracy in the world, and how democracy works, we need to be able to answer descriptive questions like these.

Causal questions

Causal questions, on the other hand, ask if some variables drive change in other variables. For example:

  • Does exposure to divisive content on social media make people more tolerant of political violence?
  • Does declining trust in political elites weaken commitment to democratic norms?
  • Do compulsory voting laws increase political participation?

Notice that in each of these questions, one thing is driving change in another thing. Exposure to divisive content is driving change in tolerance for political violence. Declining trust in political elites is driving change in the level of commitment to democratic norms. And compulsory voting laws are driving change in the level of political participation.

These questions tell us what would happen if we intervened, for instance if we were to enact a compulsory voting law. When we design a causal study, we can find out what would happen under an intervention.

Distinction, not hierarchy

This is a distinction between descriptive and causal questions, not a hierarchy. Causal questions are more directly concerned with what can be done to bring about social change, which might lead some people to believe that they are more important than descriptive questions. However, descriptive questions are a prerequisite for knowing how we can change things: a prerequisite for changing the world is understanding what it looks like. This is why causal and descriptive questions are equally important, and both deserve our attention.

Complementarities between questions

Descriptive and causal questions also complement each other. Let’s look at a series of examples that concern the same topic.

  • Describing a distribution: What percentage of the world’s wealth is held by the 50 richest people?
  • Describing a trend: Is economic inequality increasing across the world?
  • Describing a correlation: Are autocracies more unequal than democracies?
  • Probing causation: Do increases in income inequality make democratic backsliding more likely?

These are all important questions that complement each other, and that are crucial to the research agenda on the politics of economic inequality.

Recap

  • Empirical questions can be descriptive or causal.
  • Descriptive questions ask about distributions, trends, and correlations.
  • Causal questions ask about how some variables affect others: how some things drive change in other things.
  • Causal and descriptive questions complement each other to advance our scientific knowledge.

Why this matters

Descriptive and causal questions call for different research designs. We design our study, and structure our data collection and analysis, differently depending on whether we are addressing descriptive or causal questions. Knowing what type of question we’re asking is the first step towards conducting good quality research.