What is Theory?

Theory is NOT:

• Correlation
• Description
• Interpretation
• Absence of data
• Ideology

Theory IS:

• A general, abstract explanation

Theories simplify reality

• The world is complex
• Theories tell us what to pay attention to

Theories have empirical implications

• Assume the existence of a real world
• Aim to increase understanding, not to accomplish political, moral, or social ends

​

• Have empirical implications
• If X, then Y, where X and Y are observables

Social theories

• Explain social rather than individual outcomes
• We are interested in explaining the behavior of social systems rather than of individuals

​

Social theories

• Are composed of
• Causal relations
• Causal mechanisms

Causal Relations

• The outcome
• The thing you are trying to explain
• Also called
• Dependent variable
• Effect

Causal Relations, cont’d

• The cause
• A factor such that change in it is associated with change in the outcome
• Also called
• Independent variable

Causal Relations, cont’d

• So, X causes Y if by modifying X, one can affect Y
• An explanation includes an assertion of such a causal relation

Causal Mechanisms

• The link between the cause and the outcome
• In other words, the process through which the cause leads to the outcome

​

We will say more about this later

How do we know which theory is best?

• Empirical evidence
• Theories produce empirical predictions about how change in a causal variable will affect an outcome variable
• These predictions are called hypotheses

Example: Durkheim’s theory of suicide

• The level of individualism in a group affects the rate of suicide in the group
• Individualism – a cause
• The degree to which individual activities are controlled by individuals themselves rather than by others
• Suicide rate – an outcome
• Some countries/groups have a low rate; others a relatively high one

Empirical implications

• If Protestants are more individualistic than Catholics
• Then Protestants in France will have higher suicide rates than Catholics in France
• If unmarried men are more individualistic than married men
• Then unmarried men will have a higher rate of suicide than married men

Empirical implications, cont’d

• To determine whether the predictions are supported by the data, we must pay attention to three things:
• Correlation
• Causal Order
• Spuriousness

Correlation

• A change in X is associated with a change in outcome Y

Causal order

• The cause must occur before the effect
• It is possible to change the value of the dependent variable by changing the causal variable
• In other words, if you change X, Y will change

​

Cause Outcome

(x) (Y)

Non-spuriousness

• To infer causality, all possible spurious causes of Y (the dependent variable) must be ruled out
• That is, the researcher must determine that a third variable is not responsible for the observed relation between X and Y

​

Example: The Protestant Ethic

• Max Weber noted that the initial geographic distribution of European industrial capitalism seemed correlated with the % of Protestants in a country
• Protestantism 🡪 industrial capitalism
• Possible spurious causes
• Perhaps countries with large coal reserves tended to be Protestant
• If coal reserves 🡪 industrial capitalism, then Protestantism is a spurious relation

Example The Protestant Ethic

Capitalism

Protestantism

Coal Reserves

Assessing theories

• If the theoretical predictions are consistent with what we observe, then we have more confidence in the theory

Caveats

• Very few ‘classical theories’ live up to these expectations
• They do not always explicitly articulate causal relations and causal mechanisms
• Not many contemporary ones do, either

Theories are imperfect

• Theories simplify reality
• Theories must omit much, must overemphasize much
• Hence, all theories are imperfect

​

How to choose between rival theories?

• Ultimate criterion: empirical adequacy
• The best theory is the one that is most consistent with observable empirical phenomena