. . . We really do have the best comment section on the internet.

Posted by Andrew on 6 November 2016, 8:44 pm

. . . We really do have the best comment section on the internet.

## Recent Comments

- sam on What are some common but easily avoidable graphical mistakes?
- Andrew on What are some common but easily avoidable graphical mistakes?
- Nick Adams on Neural nets vs. regression models
- Jeff on What are some common but easily avoidable graphical mistakes?
- Andrew on What are some common but easily avoidable graphical mistakes?
- Matt on What are some common but easily avoidable graphical mistakes?
- Anoneuoid on Neural nets vs. regression models
- Joshua Pritikin on That illusion where you think the other side is united and your side is diverse
- Ron Kenett on Neural nets vs. regression models
- Corey on Neural nets vs. regression models
- Daniel Lakeland on Neural nets vs. regression models
- Kevin Van Horn on Neural nets vs. regression models
- Ricardo on Neural nets vs. regression models
- Tom M on Neural nets vs. regression models
- Daniel Lakeland on Neural nets vs. regression models
- Bob Carpenter on Neural nets vs. regression models
- Keith O'Rourke on Neural nets vs. regression models
- Bob Carpenter on Neural nets vs. regression models
- Charles Fisher on Neural nets vs. regression models
- Daniel Lakeland on Neural nets vs. regression models

## Categories

Research has shown that it is good.

We reject at p less than .05 the null hypothesis that our comment section is identical to all other comment sections on the internet.

To be exact, at 0.0492.

Actually .052 but through opportunistic rounding we got it down below .05 where it belonged.

P.S. Yes this blog has the best overall comment section on the internet but this particular comment thread is pretty lame (for which I accept more than 50% of the responsibility).

There must be a name for this paradox: Somebody’s Law of the Internet: Once someone starts bragging online, the brag will inevitably nullify whatever is being bragged about.

I thought the No Trump reference was 0.2% less than lame (with a 95% confidence interval of 0.1 to 0.3).

In case anyone doesn’t know what I’m referring to, here it is (I was rereading it today and cachinnating):

http://images.nymag.com/images/2/daily/2016/06/notrump_falk_gelman_icml.pdf

“When we look at the percentage of No Trump contracts made, the Vanderbilt 2015 tournament is significantly higher than the Vanderbilt 1999 tournament. The t-test yields a p-value of 0.0492, easily passing the traditional significance level of 5 percent.”

I thought the mention of this number would be Gladwellian in impact, but I don’t mind being wrong. I do sort of mind being correlated with lameness, though.

Diana:

Your comment was fine. I introduced the lameness with my followups.

I was thinking of suggesting a p.s. that many readers of this blog do not read the comments – but then this is unlikely the thread that would encourage any change in that ;-)

I participate in the best comment section on the internet! Woohoo! That’s going on my résumé.

Well, in the tradition of Stigler’s law of eponymy, here are some suggestions:

* Gelman’s law of commentary self-abrogation

* Gelmanic Reversion to the Mean

or my favorite

*Gelmanic Depression (of commentary hyperbole)

For what it’s worth, I’m not a statistician and this limits my ability to understand the more technical topics on this blog. The topics that I can follow are of interest, but one of the things that originally hooked me was the spectacle of people with strong opinions disagreeing with each other in civil fashion–on the Internet.

Also, this particular thread shouldn’t count against that appraisal. Someone once told me I received compliments gracefully and I was completely flummoxed.