Identical crime statistics can be framed to seem large or small

Evidence

Statistics have shown that by the age of 25, 96.3% of all Alphans have no criminal record whereas 98.8% of all Betans have no criminal record…. In light of the available crime statistics, which would you select? [program J, 71%; program K, 29%] … It should be apparent that the crime statistics on which respondents were to base their choice were the same across the two problems. Because of the ratio-difference principle, however, the Alphans are perceived as much more criminal than the Betans in problem 11—roughly three times as criminal—but they are seen as only slightly less noncriminal than the Betans in problem 12. As hypothesized, respondents selected that program in which differences in allocations between the groups matched as closely as possible differences in perceived criminality, resulting in a large reversal of preference. — John T. Jost and Jim Sidanius, Political Psychology, 2482 (#expands, via linked claim)

Categorization

• XREF: connects to: Criminal juries systematically convict less often than judges would in identical cases — G. Alan Tarr, Judicial Process and Judicial…, 1439; Actual crime levels run roughly three times reported UCR figures — G. Alan Tarr, Judicial Process and Judicial…, 1666; Victimless crimes and trivial thefts go unreported, skewing crime data — G. Alan Tarr, Judicial Process and Judicial…, 1669 (+4 more)

Broader Connections

• Identical crime statistics produce opposite policy preferences when reframed #expands (0.95) — John T. Jost and Jim Sidanius, Political Psychology, 2482
• Crime statistics are constructed through discursive practices, not neutral counting #complicates (0.75)