Collider Bias
Describe grandparent example
Simpson's Paradox - present in grandparent example, means including another predictor can reverse the direction of association between other predictor and the outcome. Usually the new predictor helps us, but in this one, it says the opposite.

University of Connecticut

Collider Bias - when you condition on a collider, it creates statistical - but not necessarily causal associations among its causes. 
Relationship between trustworthiness, worthiness and selection

Collider Bias

Example of happiness, age, and marriage --> shows marriage as a collider
With colliders, there should be no association between age and happiness
In model, there is a negative association but it is just statistical not causal

Learn Before

Related