Ch 5: Evidence

Attrition bias

A selection effect where patients drop out of a research study, making the evidence unreliable.

Echo chambers

A feedback loop where information sources are selected to support existing opinions, influenced by social media algorithms.

Evidence for H

A fact that is more probable given hypothesis H than not-H, increasing confidence in H.

Evidence test

A method to determine if a new fact supports hypothesis H by comparing its likelihood under H versus not-H.

File-drawer effect

A selection effect where researchers do not publish studies unlikely to be accepted, leading to publication bias.

Heads I win, tails we're even

An error in evaluating evidence where new facts against a favored position are ignored.

Hypothesis

Any claim under investigation, often represented by the letter "H."

Independent of H

Refers to evidence that does not affect the hypothesis H, as per the evidence test.

Media bias

The tendency of media to present content that engages viewers, potentially reflecting political biases.

One-sided strength testing

Evaluating evidence only based on its likelihood if a claim is true, ignoring its likelihood if false.

Opposite evidence rule

A guideline to consider how we would react to evidence against our view to avoid bias.

Publication bias

The tendency for journals to publish surprising research while overlooking studies that support conventional wisdom.

Selection effect

A factor that systematically influences what can be observed, potentially leading to unreliable evidence.

Selective noticing

The tendency to notice evidence supporting a hypothesis while ignoring disconfirming evidence.

Serial position effect

The tendency to remember the first and last items in a series better than those in the middle.

Strength factor

A measure of how strong evidence is for hypothesis H, calculated by comparing probabilities under H and not-H.

Strength test

A comparative evaluation of evidence's likelihood under hypothesis H versus not-H to determine its strength.

Survivor bias

A specific selection effect where only surviving information sources are considered, leading to skewed conclusions.