Before Citing Anything, I Now Ask Scite One Question
Scite shows which later papers support or contradict a claim. That one check keeps me from citing something walked back.
I used to cite a paper because it said what I wanted. Then Scite showed me a study I loved had been contradicted three times since. Now, before any citation, I check its support-contrast count in Scite.
The question is simple: 'does the field still stand behind this, or has it moved?' A claim with ten supports and one contrast is safe to cite. One with a near-even split is a debate I should name, not a fact I should state.
It also catches the quiet retraction. A paper can be widely cited and later weakened, and Scite surfaces the contradiction before I embed the claim in something a reviewer reads. That catch has saved me twice from looking dated.
Scite isn't a replacement for reading the paper; it's the pre-read that tells me which papers to read closely. Five seconds per citation, and the ones I'd have been embarrassed by get flagged first.