hello

@AndyW Stephen Stigler published one in one of his books. Something similar is reproduced in his slides at www.statoo.ch/sst07/presentations/Stigler.pdf; see p. 10. It shows the first one (1873).

@whuber: Sorry for bothering you here with this, but I would be very interested in hearing your opinion regarding my recent question on meta about merging duplicates: meta.stats.stackexchange.com/questions/2310. I asked there "what our policy is", but actually my point was rather that we should merge more. I would be grateful if you could answer there. In fact, I would be happy to hear from other moderators as well, as you are the only few people who can make decisions on merging.

I have a question on writing up Welch t-test for publication with a bounty on it, that risks slipping into unclaimed territory! stats.stackexchange.com/questions/124961/… I wonder if it might be because there is an answer already so people assume it's all done and dusted.

I am a bit of a loss on how to stimulate more interest. The question would be well-suited for users who regularly have to write up such results, and have an awareness of appropriate style guides or journal requirements (I mentioned APA as a particular example). There are probably plenty of folk on the Cognitive Science or similar SEs who would know the answer off-hand!

Any suggestions for a suitable way to stimulate interest among relevant users would be appreciated.

@Silverfish Isn't there already guidance on reporting p-values? That would imply (somewhat indirectly) how df should be reported: use as much precision as needed to reproduce the reported p-value.