Post Hoc & Ad Hoc

Views:
 
Category: Education
     
 

Presentation Description

Scientific thought and the problems of post hoc and ad hoc explanations.

Comments

Presentation Transcript

Post Hoc & Ad Hoc : 

Post Hoc & Ad Hoc headlessprofessor

Post Hoc & Ad Hoc : 

Post Hoc & Ad Hoc Explanations as to why: causal speculation

Post Hoc & Ad Hoc : 

Post Hoc & Ad Hoc Explanations as to why: causal speculation Both are flawed science

How to infer cause & effect : 

How to infer cause & effect Independent Dependent X Y cause effect

Relationship between variables : 

Relationship between variables X caused Y

Relationship between variables : 

Relationship between variables X caused Y Y caused X

Relationship between variables : 

Relationship between variables X caused Y Y caused X Both are effects

Relationship between variables : 

Relationship between variables X caused Y Y caused X Both are effects (spurious)

Post hoc fallacy : 

Post hoc fallacy Spurious correlation = collateral effects due to same underlying cause

Spurious Relationship : 

Spurious Relationship X Y Z

Relationship between variables : 

Relationship between variables X caused Y Y caused X Both are effects Null hypothesis

How do we prove that : 

How do we prove that X caused Y ?

How to infer cause & effect : 

How to infer cause & effect Reject the null to establish significance

How to infer cause & effect : 

How to infer cause & effect Reject the null to establish significance p < .05

How to infer cause & effect : 

How to infer cause & effect Reject the null to establish significance p < .05 inferential statistics

How to infer cause & effect : 

How to infer cause & effect Reject the null to establish significance Identify the direction of the correlation

How to infer cause & effect : 

How to infer cause & effect Reject the null to establish significance Identify the direction of the correlation If X helps Y, the correlation should be positive

How to infer cause & effect : 

How to infer cause & effect Reject the null to establish significance Identify the direction of the correlation If X hurts Y, the correlation should be negative

How to infer cause & effect : 

How to infer cause & effect Reject the null to establish significance Identify the direction of the correlation Identify if one variable is the IV

Experiments are the best way to : 

Experiments are the best way to confirm a causal relationship

Fallacies of scientific induction : 

Fallacies of scientific induction Post hoc ergo propter hoc

Fallacies of scientific induction : 

Fallacies of scientific induction Post hoc ergo propter hoc fallacy of causal reasoning

Fallacies of scientific induction : 

Fallacies of scientific induction Post hoc ergo propter hoc fallacy of causal reasoning assumes that an event which occurred before the effect must be its cause

Post Hoc jumps to this : 

Post Hoc jumps to this X caused Y

Post Hoc ignores Spurious : 

Post Hoc ignores Spurious X Y Z

Example of Post Hoc : 

Example of Post Hoc Of all the different religions in the U.S., Hindus have the highest median income

Example of Post Hoc : 

Example of Post Hoc Of all the different religions in the U.S., Hindus have the highest median income Causal inference: being a Hindu helps you get rich

Alternative Explanation : 

Alternative Explanation Many Hindu immigrants to the U.S. are investors, entrepreneurs or professionals

Alternative Explanation : 

Alternative Explanation They are not wealth because they are Hindus

Alternative Explanation : 

Alternative Explanation They are Hindus because they come from India

Alternative Explanation : 

Alternative Explanation They are Hindus because they come from India They are wealthy because of selective migration from India

Spurious correlation : 

Spurious correlation X Y Z

Alternative Explanation : 

Alternative Explanation Hindu Wealthy X Y Z Educated Immigrant

Post hoc fallacy : 

Post hoc fallacy correlation: association between two variables

Post hoc fallacy : 

Post hoc fallacy correlation: association between two variables permits prediction, but not always causal inferences

Post hoc fallacy : 

Post hoc fallacy correlation: association between two variables permits prediction, but not always causal inferences Experiments get around post hoc

Ad hoc hypothesis : 

Ad hoc hypothesis An explanation flexible enough for all occurrences

Ad hoc hypothesis : 

Ad hoc hypothesis An explanation flexible enough for all occurrences Cannot be rejected by any empirical data

Karl Popper : 

Karl Popper “A hypothesis is only scientific if it could be rejected by empirical evidence”

Religious Claim : 

Religious Claim God is all powerful, all wise and always wants what is best for us.

Religious Claim : 

Religious Claim God is all powerful, all wise and always wants what is best for us. If I am healthy and prosperous, thank God.

Religious Claim : 

Religious Claim God is all powerful, all wise and always wants what is best for us. But, if I am suffering …

Religious Claim : 

Religious Claim God is all powerful, all wise and always wants what is best for us. suffering… Ad hoc explanation: we cannot understand God’s wisdom.

Purposes of Science : 

Purposes of Science Understand (causal analysis) Predict (testable hypotheses) Control (technology)

Psychoanalysis : 

Psychoanalysis Freud’s theory that unconscious sexual & aggressive drives determine human behavior

Psychoanalysis : 

Psychoanalysis If someone acts out Freud’s theory explains that.

Psychoanalysis : 

Psychoanalysis If someone does not act out Freud’s theory explains that.

Psychoanalysis : 

Psychoanalysis explains anything.

Psychoanalysis : 

Psychoanalysis explains anything. predicts nothing.

Psychoanalysis : 

Psychoanalysis Popper: “psychoanalytic explanations are ad hoc”

Both post hoc and ad hoc : 

Both post hoc and ad hoc Frustrate causal analysis in science.

Both post hoc and ad hoc : 

Both post hoc and ad hoc Frustrate causal analysis in science. Refuse to consider an alternative explanation

Solution : 

Solution Standards of scientific methodology

Solution : 

Solution Standards of scientific methodology Peer review

Post Hoc & Ad Hoc : 

Post Hoc & Ad Hoc headlessprofessor

authorStream Live Help