2 ms·
There's no way to guard against all false positives. And while changing the P value cutoff would reduce them, it would also increase false negatives. The answe
by ineedasername 7y ago
There's no way to guard against all false positives. And while changing the P value cutoff would reduce them, it would also increase false negatives.
The answer doesn't lie in hard-line stances for or against P values or with an alternative that will have its own set of problems. It lies with greater education of those who run experiments & those who consumer the literature about proper interpretation and other methods of analysis that should be used along side it.
- merpnderp 7y agoOther small things could help, like publishing failures, incentivizing duplication of experiments, increasing access to results.
- ineedasername 7y agoI think duplication and access to raw data would be two huge steps forward. It's actually a bit mystifying that researchers are able to present their own final conclusions & interpretations without making the actual data available.