Choose a favorable sample
Exclude inconvenient data
Use misleading averages
Change the time frame
Cherry-pick comparisons
Confuse correlation with causation
Use vague or undefined terms
Hide the base rate
Present percentages without raw numbers
Use small sample sizes
Truncate axes on graphs
Use inconsistent scales
Group data selectively
Ignore outliers when convenient
Overstate precision
Use technical jargon to obscure meaning
Report relative change instead of absolute change
Omit uncertainty and error margins
Frame results with loaded language
Repeat claims until they seem true