Probability and Statistical Biases
Statistical biases distort how we read evidence, samples, studies, and patterns in data.
01
Base rate fallacy
Ignoring general probabilities when judging a specific case.
02
Selection bias
Drawing conclusions from a non-representative sample.
03
Survivorship bias
Studying only those who made it through a process.
04
Publication bias
Published evidence overrepresents notable or positive findings.
05
Reporting bias
Only some outcomes or facts are reported.
06
Sampling bias
The sample differs systematically from the population.
07
Observer bias
Observer expectations influence measurement.
08
Confirmation bias
Data is interpreted to support an existing belief.
09
Measurement bias
The measurement tool systematically distorts results.
10
Attrition bias
Dropouts from a study change the results.
11
Recall bias
People remember past events inaccurately or unevenly.
12
Nonresponse bias
People who do not respond differ from those who do.
13
Lead-time bias
Earlier detection seems to improve survival without changing outcome.
14
Length-time bias
Slower cases are more likely to be detected in screening.
15
Ecological fallacy
Assuming group-level data applies to individuals.
16
Modifiable areal unit problem
Results change when geographic boundaries change.
17
Multiple comparisons problem
Testing many relationships makes false positives more likely.
18
P-hacking
Trying analyses until a significant result appears.
19
Texas sharpshooter fallacy
Drawing the target around data after seeing it.
20
Regression to the mean misunderstanding
Mistaking natural return toward average for a causal effect.