Lecture 016

Objection:

Reading

Democratic Peace Theory

democratic peace: democratic countries do not fight each other

Reason for causation: we want to control saturation

Critiques:

Law and Correlation

Difference between law and correlation:

Creativity

Do (causal) social law exists?: it exists if one believes social science theory are generalizable. When human have "creativity"

Do human have creativity

Social laws can be formulated in another vocabulary (like physical science)

Is human predictable: Consider chaos theory

Causality

Most famous definition

Hume's causality: an object followed by another, and where all the objects, similar to the first, are followed by objects similar to the second. Or, in other words, where, if the first object had not been, the second never had existed.

Critiques:

Probabilistic dependency

There is no causality in linear regression

Interventionism

Thomas Reid's causality: controlled change of variable limitation

adjusted definition: A necessary and sufficient condition for X to be a direct cause of Y with respect to some variable set V is that there be a possible intervention on X that will change Y (or the probability distribution of Y ) when all other variables in V besides X and Y are held fixed at some value by interventions.

Two advantage

Questions

  1. policy recommendations need causal claims to justify. (Justifying spreading of Democracy). The need for causation is because we want to control things.

  2. interventionist vs. "regularity account"

  3. a natural law: in its definition, should be context-less

  4. interventionist can rely on exact context to produce "imagined" outcome Since we can't "observe causality", we need to establish a standard to decide whether a causal statement has "causality credibility"

  5. regularity account: gains its credibility by generalization

  6. interventionist: gains its credibility by interventionist

  7. Justification for linear model?

  8. it might be non-linear

- it might not be caused

  1. Firstly, this linear equation model disregards the correlation coefficient, making the equation a non-probabilistic model. In a non-probabilistic model, the model seems to suggest that if I plug variables (my years of education, my score on the last state exit exam, etc.) into the equation, I expect to get the exact result (the exact number of my income). Since this is rarely the case, the equation is a bad explanation of influences on income. Secondly, the linear equation model seems to suggest a correlation than causation because the model is likely created by individual data points collected from uncontrolled observations. One might interpret this linear equation as one spends more years in an educational system causes the person to receive a higher income. But this causality does not necessarily hold because one can argue that a person with a higher income allows the person to spend more years in the educational system. Thirdly, even if causality in the study is supported by good experimental practice using interventionist methods, one who believes in human's free-well may argue that the income of one person can't be caused by variables like "Years of Education" or "Score on last state exit exam take" since it is ultimately the boss's psychology that makes his act of giving a person certain money.

Class

Regularity: let X in set XS, Y in set YS, for all X and Y...

Interventionist: a definition in terms of intervention

Table of Content