1.5 Empirical Political Science

Learning Outcomes

By the end of this section, you will be able to:

• Distinguish empirical political science from normative political science.
• Explain what facts are and why they may be disputed.
• Define generalization and discuss when generalizations can be helpful.

Unlike normative political science, empirical political science is based not on what should be, but on what is. It seeks to describe the real world of politics, distinguishing between what is predictable and what is idiosyncratic. Empirical political science attempts to explain and predict.³²

Empirical political science assumes that facts exist: actual, genuine, verifiable facts. Empirical questions are ones that can be answered by factual evidence. The number of votes a candidate receives is an empirical matter: votes can be counted. Counting votes accurately so that each candidate receives the actual number of votes that were cast for them can be difficult. Different ways of counting can lead to slightly different counts, but a correct number actually exists.

A fact may be disputed. There may be genuine uncertainty as to what the facts really are—what the evidence really shows. Sometimes it is extremely difficult to gather the facts. Do space aliens exist? That is an empirical question. Either space aliens exist, or they do not. Some researchers claim to have evidence that space aliens are real, but their evidence is not universally, or even broadly, accepted. One side of this argument is correct, however, and the other is not. Evidence has not yet conclusively determined which is correct.³⁵

Does the Russian government seek to interfere with American elections, and if so, does its interference affect the outcome? The first part of the question is difficult (but not impossible) to answer because when a country interferes in another country’s domestic affairs it tries to do so in secret. It is difficult to uncover secrets.³⁶ But the second part of the question, does the interference affect the outcome, is almost impossible to answer. Because so many factors influence election outcomes, it is extremely challenging to determine which individual factors made any consequential difference.³⁷

There are thus empirical debates in which people of good faith disagree about what the facts are. In many cases, however, people do not want to acknowledge what the evidence shows, and because they do not want to believe what the facts demonstrate, they insist the evidence cannot be true. Humans often use motivated reasoning, first deciding what is true—for example, “Gun control makes us safer” or “Gun control makes us less safe”—and then finding evidence that supports this belief while rejecting data that contradicts it.³⁸

In other cases, individuals and interests may actually know what the facts are, but they are motivated by reasons of self-interest to deny them. The evidence is clear, for example, that nicotine is addictive and harmful to human health. The evidence is also clear that Big Tobacco, the largest cigarette companies, denied these facts for years because to admit them would have put their profits at risk.³⁹

Former President Donald Trump, along with many of his supporters, claims that he won the 2020 presidential election and that President Joe Biden was declared the victor only because of massive voter fraud. All attempts to prove that fraud led to Biden’s victory have failed: no evidence has been found to support Trump’s claims.⁴⁰ That these claims continue can be attributed to the fact that some individuals are simply unwilling to accept the evidence, while others benefit from denying the validity of it.⁴¹

Empirical political science might find—based on the available evidence—that individuals with more education or more income are more likely to vote. Empirical political science would not consider whether this is good or bad; that would be a normative judgement. Empirical political scientists might explain the link between education, income, and voting by positing that better educated, more prosperous individuals are more likely to believe that their views matter and that because of that belief they are more likely to express those views at the ballot box. These political scientists might also use their findings to make a prediction: an individual with more education or higher income is more likely to vote than an individual with less education or lower income.⁴²

Based on this finding, empirical political scientists make no claims as to who should participate in politics. Questions about “should” are the domain of normative political science. Moral judgments cannot be made strictly on the basis of empirical statements. That members of one group vote at higher rates than another group, for example, tells us nothing about whether they deserve to vote at higher rates or whether government policies should be based more on their views as compared to those who vote at lower rates.

From this finding, however, empirical political scientists may infer a generalization. Generalizations are based on typical cases, average results, and general findings. Younger adults, for instance, typically vote less often than older adults. This does not mean that any specific young adult does not vote or that any specific older adult does, but that these statements are generally true.⁴³

Generalizations can be helpful in describing, explaining, or predicting, but there is a downside to generalizations: stereotyping. If the evidence shows that political conservatives in the United States are opposed to higher levels of immigration, this means neither that every conservative holds this belief nor that one must hold this belief to be conservative. If data suggests supporters of abortion rights tend to be women, it is not possible to infer from the evidence that all women seek more permissive abortion laws or that no men do. In using generalizations, it is important to remember that they are descriptive of groups, not individuals. These are empirical statements, not normative ones: they cannot by themselves be used to assign blame or credit.

Empirical political science can be used to make predictions, but predictions are prone to error. Can political science knowledge be useful for predicting the outcome of elections, for example? Yes. Given a set of rules about who is eligible to vote, how votes can be cast, and what different categories of voters believe about the candidates or policy options on the ballot, political science knowledge can be useful in predicting the outcome of the election. Our predictions might be wrong. Maybe people did not tell the truth about who they were planning to vote for. Maybe the people who said they were going to vote did not.

In 2016, most political polls predicted that Hillary Clinton would be elected president of the United States.⁴⁴ Clinton did indeed win the popular vote, as the pollsters anticipated, but Donald Trump won the electoral vote, against the pollsters’ expectations. Political science is imperfect, but it seeks to learn from and correct its mistakes. You will learn more about public opinion polling in Chapter 5: Political Participation and Public Opinion.

Many of the terms in this book, like incumbent, are relevant mainly for the study of politics. Other terms, like ceteris paribus, are useful across a broad range of studies that use the scientific method. Ceteris paribus can be translated as “all other things being equal.” If the ethnicity of a political candidate does not influence their probability of getting elected to office, ceteris paribus, if there are only two candidates and if they are alike in every relevant aspect (e.g., age, experience, ability to raise campaign funding) except their ethnicity, then the candidate’s ethnicity by itself does not affect the outcome of the election.

In real life, however, “all other things” are almost never equal. To the extent that our societies have inequalities of wealth, health, education, and other resources, the inequalities tend to be correlated—that is, mutually related—to each other. For example, wealth and health are correlated with each other in that wealthier people tend to have better health and poorer individuals tend to have poorer health. In the United States, Whites tend on average to have more wealth, health, education, and other social resources than do persons of color.⁴⁵ This does not mean that every White person is wealthier and healthier, but that on average, in general, they tend to be.

Empirical political science and political philosophy (or normative political science) are distinct modes of inquiry. But this is not to say that they are conflicting, that one is better than the other, or that political scientists do not use both in their research. If empirical research discovers that certain groups are systematically disadvantaged in the political process, the researchers may also argue that these disadvantages are harmful or wrong and make a moral argument that the disadvantages should be reduced or eliminated. Empirical research is often inspired by normative concerns. Those who believe that human rights should be better protected may undertake research to understand the political factors that limit the protection of rights.