**Source identification:** 2019 meta-analysis on caffeine and cognitive performance; lead author at public university; partial funding from beverage industry association. **What the source claims:** Caffeine moderately improves working memory and attention in healthy adults, with effect sizes in the small-to-medium range across 47 included studies. **Methodology read:** Meta-analytic; aggregates 47 prior studies. Likely pre-registered if peer-reviewed in a reputable venue; would need to verify. Standard meta-analytic concerns apply: selection of studies, handling of heterogeneity, publication bias correction. **Strength signals:** - Aggregating 47 studies gives more statistical power than any individual study. - The "healthy adults" framing limits the population, which reduces some confounders. - The claimed effect sizes are modest, which is consistent with what the existing literature would predict — a wildly large claimed effect from a meta-analysis on a well-studied topic would be more suspect. **Weakness signals:** - Industry funding for caffeine research is a real conflict-of-interest concern. Meta-analyses of caffeine consistently show modest positive effects, but the magnitude can shift depending on whether industry funding correlates with study inclusion criteria. - "Working memory" and "attention" are operationalized differently across the underlying studies. The meta-analytic effect size is an average across heterogeneous outcomes, which can obscure when the effect is real for one type of task but not another. - Without access to the protocol, can't tell whether they corrected for publication bias appropriately. Caffeine research has a known file-drawer problem. **Where I'd downweight specific claims:** - The "attention" claim. Attention measures vary widely (sustained vs. divided vs. selective); a single meta-analytic estimate hides the variance. - Any claim about clinical significance. A small-to-medium effect on a cognitive task is not the same as a small-to-medium effect on real-world performance. **Where I'd uprate this source:** - The fact that the conclusions are modest (no claims about caffeine being "the secret to genius") suggests the authors aren't overstating. Sources that match the strength of their evidence to the strength of their claims deserve more weight, all else equal. **Bottom line:** Moderate weight. Useful as evidence for "caffeine probably helps cognitive performance somewhat" but not as evidence for any specific effect size in any specific population or task. The industry funding is a real concern but not disqualifying; the modesty of the claims partially offsets it.