Thread by @carabonate

Many journals from @theAGU now display a @scite badge above each article, which claims to quantify how many times subsequent papers 'dispute' and 'support' the article's results.

I'm concerned that the @scite algorithm misidentifies which works are disputed. #AGUpubs 1/

In environmental science, we frequently compare our results with data from another time/place. We perform a literature review of related datasets, and may state that our results "contrast with Zhao et al." @scite often flags this type of phrasing as 'disputing' Zhao's results. 2/

Here's an example:
"However, we have not found any correlation between CH4 and MeHg in surface waters, unlike other studies (ref A), possibly due to CH4 being oxidized in CO2 during its transport from the deep parts of the water column to surface waters (ref B, ref C). " 3/

.@scite says this text disputes refs A, B, and C (with 99, 99, and 94% confidence, respectively).

But the authors used refs B & C to explain why their results, in deeper waters, may be different from ref A, which was in shallow waters. None of these works are being disputed! 4/

Should I stop citing related works in my papers so that @scite won't involve me and my coauthors in imaginary disputes?

Should I stop publishing papers on the rapidly-changing Arctic because they may receive a high @scite 'dispute' score when others publish new data? 5/

.@theAGU please reconsider your participation in the @scite badge pilot program with @wileyinresearch/@WileyGlobal.

I am uncomfortable that an inaccurate metric of how many citations have disputed and supported the work appears before the article text in #AGUpubs. 6/6