Interesting (non-research) Statistics Articles


Can't make spagetti
amazing stuff. I went to Disney World last November to celebrate my husband's birthday and I got one of those bands. AMAZING STUFF! the future is noooowww!!!


Can't make spagetti
in my field (social sciences/psych) there's been a sort of "mini-revolution" towards trying to make science more transparent and encourage data-sharing. which makes sense in the light of big-time names being discovered of committing fraud (Diederik Stapel anyone?

being the glass-half-empty, cynical, somewhat dark-mood person that I am, I always thought this idea of a "transparency in psychology" was good in terms of intentions but that its implementation may prove to be so cumbersome that people would probably end up just abandoning the idea all together.

well... as I read new the blog entry of one of my favourite psychologists/philosophers of science, I can see that the road ahead for this "transparency" is gonna be a lot steeper than I anticipated, since the opposition is not only coming from the authors of research themselves but also from the organisation that administers the journal:


Super Moderator
That story of Denny's is a bit of a shocker. Nice to see that he resigned from the journal in protest.

I think applying open data policies in psychology (or human sciences generally) is going to be a hard road though. E.g., at my university the code of ethics says that data can't be re-used in a new research project without re-contacting the participants and getting their informed consent. I'm not sure exactly what counts as "new research" in this context, but this kind of idiotic restriction makes it pretty **** hard to share data without potentially resulting in a breach of the ethics guidelines.
Last edited:


Less is more. Stay pure. Stay poor.
The big stance is if the gov pays for the research they want these data out there to get more bang for their buck.

Dason, good idea.


Ambassador to the humans
The thread isn't quite popular enough to warrant it yet but I'm going to sticky the thread so it's easier to find later.
I saw on statblogs that Rob Hyndman had commented on an article by Hal Varian. It is interesting to see Hyndmans evaluation.

Everybody loves Varian (Googles chief economist) because he had said that "the future belongs to the statisticians". Varian's article is freely available here. The paper seems to me to be a nice summary of some relatively recent methods.

I guess that the the title "Big Data: New Tricks for Econometrics" alludes to the story about the econometric professor, who got a new secretary that instead of typing "econometrics" typed "economic tricks".

Varian is referring to a textbook by James Witten Hastie and Tibshirani "An Introduction to Statistical Learning with applications in R" (2014). The text is freely available here. So they continue the great "tradition" in not only publishing it in a book, they also make it freely available. (Please post here if you see a "second printing".)
Research suggests that it depends on your initial level of confidence after getting the directions. Did you hear them right? Did you turn at the 3rd light? Could you have driven past the restaurant? Is it possible the directions are incorrect?