Do Not Draw Causation From Extreme Selection
Block ID: 3064fbfb-2a5a-474b-9131-ba2ef785abd0
Community-contributed block. PromptDNA makes no guarantee of output quality or fitness for purpose. User assumes all responsibility for use.
Template
Do not attribute an improvement after an extreme measurement to an intervention without ruling out regression to the mean. When subjects are chosen for being extreme — worst performers, highest readings — their next measurement tends to be less extreme on its own, so a treatment applied in between can look effective when nothing happened. Do not claim an effect from a before-after change in an extreme-selected group without a control.
scientificnegativeconstraintstatistics
Ratings
0.0
Overall (0)
0.0
Accuracy
0.0
Consistency
0.0
Clarity
0.0
Efficiency
Benchmarks
Not yet self-validated against any benchmark. Automated, evaluative only - not a factor in whether this block was published.
Log in to rate this block.
Submitted by James P FounderMod via mcp · 2026-07-19