4 ms·
Performing studies with low N makes it more likely for spurious effects to pass the various statistical effects. Therefore, it makes it more likely to obtain a
by nhaehnle 9y ago
Performing studies with low N makes it more likely for spurious effects to pass the various statistical effects. Therefore, it makes it more likely to obtain a result that can be published (despite being wrong).
- criddell 9y agoI would think that a high n study finding no relationship between telomere length and diet would be very publishable.
- nhaehnle 9y agoTraditionally not really, unfortunately. The problem is that negative results usually don't allow you to tell a story. The solution to this is a publication process where you register studies before you do them, in exchange for guaranteed publication once you're done with them. Different fields are further along in establishing such processes. (I've never worked in an empirical field myself, just reporting what I hear from a psychologist friend.)