liveBook · Manning

Confirmation bias

As explained in Grokking Statistics

The tendency to favor, seek, or interpret evidence in ways that support an existing belief or desired conclusion.

Confirmation bias is the tendency to gather information that supports a desired conclusion while dismissing evidence that challenges it. It can steer research, data selection, and interpretation toward a preferred story. section 9.7 section 1.4

The UFO video

After seeing strange lights and black SUVs near a state park, a UFO enthusiast treats the scene as evidence of a government cover-up. When a neighbor suggests that a movie crew caused the lights, he dismisses the message as hearsay and uploads the preferred explanation anyway. The alternative explanation was available, but an existing belief and personal investment made it easy to reject. section 9.7

Reduce confirmation bias by changing how you investigate:

  • Notice when you are selecting only supportive information, then stop and question the selection. section 9.7
  • Actively allow alternative explanations, including explanations that are inconvenient or not self-serving. section 9.7
  • Look for contrary findings instead of cherry-picking results after repeatedly observing or modeling the data. section 9.16
  • Watch for organizational incentives that publish favorable studies while burying unfavorable ones. section 9.7

How Grokking Statistics covers it

How deep it goes
Defines the bias, illustrates its reasoning pattern, and gives recognition and mitigation steps without formal measurement.
Examples
Uses a UFO cover-up claim, selectively published organizational studies, and cherry-picked modeling results as examples.
What it assumes
Assumes an existing belief and competing evidence, then defines how selective gathering and dismissal distort judgment.
What you can do after
Practice/perform: recognize selective evidence, test alternative explanations, and seek contrary findings before accepting a conclusion.
How it connects
Connects confirmation bias with cognitive bias, anchor bias, p-hacking, data bias, and the sunk cost fallacy.

Also explained in

Compare how 4 books explain confirmation bias →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Grokking Statistics. The section links point to the parts of the book it draws on.