Politics · Oct 17

The campaign says crime fell 40 percent, but the dates stop before last summer

Politics

A mayoral mailer landed on my porch with “CRIME DOWN 40%” across the front. In tiny print, the comparison was between one six-month period three years ago and the same months the following year. The chart stopped before a heavily discussed rise the next summer.

The candidate's campaign says the dates match the first policy year and use police data. An opponent calls the number fabricated. I downloaded the cited table. The arithmetic appears to match the selected categories and dates, but that does not settle whether the comparison represents the broader claim on the front.

I would describe the number as selected, not automatically false. Useful questions are why those months were chosen, which offenses were included, whether reporting practices changed, and what the same measure shows across a longer period. The police dashboard and campaign mailer may also define “crime” differently.

The biggest type on the postcard makes a broad claim. The smallest type makes a narrow one. Voters should not have to discover the difference with a ruler and a spreadsheet.

Discussion

4 community replies

@TheoInTransit

Recreate the campaign’s calculation first, then extend the same categories across adjacent periods. That shows whether the dispute is arithmetic, definition, or cherry-picked timing.

@softfocus

A first-year comparison can be a legitimate way to discuss an administration’s early record. The front of the mailer should still name that period instead of implying a timeless citywide result.

@MorganSays

Yes—the correction is not “never use a bounded comparison.” It is to put the boundary where ordinary readers can see it.

@JulesAfterDark

Did the underlying dashboard change offense categories or reporting coverage during the period? A clean percentage can hide a broken series.

The campaign says crime fell 40 percent, but the dates stop before last summer

Discussion

More in Politics

Protest marshals were told to photograph anyone who “looked like trouble”

A local podcast offered candidates a paid “interview package” without telling listeners

A campaign volunteer cheered when challengers targeted every ballot from one precinct

The mayor’s public security detail stayed for the campaign fundraiser