Important or interesting research done in your field?

June 2, 2012 7:22 PM

What is some recent important or interesting research done in your field?

I am a statistician teaching a university statistics course. In the course, we try to use examples from all fields to show how statistics can be used. I regularly read the newspaper looking out for any studies. I also scan journals of all types: engineering, economics, education, political science, ecology and medical research to name a few.

However, because I am not in those fields, I'm not sure what the most interesting or important research has been. Also, Recent varies for different fields so don't think your idea is too old or too boring. I'm sure there are really cool archaeology studies I've never heard of and my students have looked at some really boring examples. Please give me some new ideas!

posted by benthegirl to Education (17 answers total)

What level stats? What types of tests/analyses do you want to focus on?
posted by griseus at 7:40 PM on June 2, 2012

Best answer: Well, since you've mentioned archaeology, surely there are some interesting ones in this bibliography, or perhaps this, or this?

Personally I think the British coin hoard ones are beyond fascinating, but I recognize that is very untrue for most people, so I tried to find some variety...
posted by jetlagaddict at 7:51 PM on June 2, 2012 [1 favorite]

Recent advances in wind generation forecasting (summary, pointers therein) have allowed vastly better understanding of wind power in transmission system integration.
posted by scruss at 8:01 PM on June 2, 2012

Response by poster: griseus: What level stats? What types of tests/analyses do you want to focus on?

In order: Any kind of t test/ANOVA, Any test done on a table (contingency table), correlation (both parametric and non parametric), regression and multiple regression, ordinal data, other non parametric methods, and GLMs. Spatial and time series, data driven methods, and - not so much, but could be interesting for a different time or class.

One example of a paper I found was "Effect of Calcium Chloride on the Quality of Canned Bartlett Pears" from Advances in Food Science. But we also looked recently at several medical studies. One on categorizing faces and one on low birthweight babies.
posted by benthegirl at 8:02 PM on June 2, 2012

Here's a fun one - this is the paper in which a group at CERN seemed to have measured a faster-than-light neutrino velocity.
posted by Salvor Hardin at 8:21 PM on June 2, 2012

Best answer: I remember reading this paper a while ago on Gene Expression Profiling and Clinical Outcomes of Breast Cancer and thought it was great....but I was looking at it from a perspective of teaching oncology, not statistics, so I'm not sure how well it would work for you.
posted by Lt. Bunny Wigglesworth at 8:24 PM on June 2, 2012

Well, it's not exactly new or empirically interesting, but the 2009 "voodoo correlations" paper by Vul, et al., made quite a splash in cognitive neuroscience. It's a highly critical methods paper that highlights some difficulties in the analysis of MRI data and various fixes for the multiple comparisons problem, both effective and ineffective.
posted by Nomyte at 8:47 PM on June 2, 2012

Best answer: There's a lot of interesting new stuff going on in digital divide research. Here's an example from the Journal of Computer-Mediated Communication: Number Matters: The Multimodality of Internet Use as an Indicator of the Digital Inequalities. As a bonus, JCMC is an open-access journal.
posted by DiscourseMarker at 9:28 PM on June 2, 2012

Best answer: There's a really fun quantitative linguistics study that shows that people hear vowels that are more similar to New Zealand vowels if they are primed by seeing a stuffed toy kiwi right before the study, and they hear the exact same vowels as being more like an Australian accent if they saw a stuffed toy kangaroo instead. AND the effect is more pronounced in rugby fans.

Hay, J. and Drager, K. (2010) Stuffed toys and speech perception. Linguistics 48(4): 865-892.
posted by lollusc at 9:41 PM on June 2, 2012 [5 favorites]

Oh, and these are New Zealanders who are hearing the vowels, so it's not like they don't know what a NZ or Australian accent sounds like. And they even claim in debriefing afterwards that they knew what the accent was (and get it correct) - they just identify the specific sound as closer to the Australian or New Zealand accent than it actually is.
posted by lollusc at 9:43 PM on June 2, 2012

Here is my favorite article published recently in my field. It is an article that looks at how long a firm needs to sustain high performance, before we can be certain that this performance is due to a competitive advantage rather than chance.

Henderson, A.D., Raynor, M.E. & Ahmed M. 2012. How long must a firm be great to rule out chance? Benchmarking sustained superior performance without being fooled by randomness. Strategic Management Journal 33(4): 387-406.
posted by bove at 9:50 PM on June 2, 2012

I like to use examples of research on plagiarism since it's always relevant, and using it as a framework can help reinforce the importance of ethical academic practice. Here's an example from bioinformatics that utilizes somewhat complicated probabilistic models, but also uses the idea of specificity and sensitivity, which might be useful for your class. And here's someone's master's thesis that looks at "Counterproductive Student Behavior" and appears to include both basic and more complicated statistical techniques.
posted by gubenuj at 10:40 PM on June 2, 2012

You might find the multilevel-regression-and-poststratification stuff from political science interesting as a way to estimate (group) level opinion with sparse data. Like, estimating state opinion from a single national poll. See Lax and Phillips 2009-now or, IIRC, Gelman, Park and Bafumi from the earlier 00s.
posted by ROU_Xenophobe at 12:00 AM on June 3, 2012

It's a staple of high energy particle physics.

'Remember that “3 sigma” is the minimum standard required for physicists to take a new result at all seriously; if you want to get really excited, you should wait for 5 sigma significance.'
http://blogs.discovermagazine.com/cosmicvariance/2011/11/14/new-physics-at-lhc-an-anomaly-in-cp-violation/

Higgs signal grew from 3.8 to 4.3 sigma

Pulling back to the wide angle view, 'what's the context here, what does this mean?' we're trying to understand why the universe is mostly matter (and not evenly matter and antimatter) or why stuff has mass. And what we think is true depends on statistics.

Turning from the universe to war:
Lancet Survey of of Iraq War casualties
There's also the german tank problem:
http://en.wikipedia.org/wiki/German_tank_problem
posted by sebastienbailard at 12:01 AM on June 3, 2012

I think this is an interesting teaching example, since all you need to understand it is randomization, means, variances. Similarly the controversy over whether Mendel's data was partially fake. An important, kind of obvious in retrospect, and extremely annoying result was this one showing that you can pick out individuals using just allele frequencies from a GWAS (you can actually sometimes do this with just per-SNP p-values!). Similar results from the Netflix Prize are more complicated.
posted by a robot made out of meat at 7:26 AM on June 3, 2012

The voodoo correlations paper Nomyte mentioned is a good example, but kind of dry. One that's maybe a bit more interesting is where researchers reported brain activity in a dead salmon.
posted by logicpunk at 12:56 PM on June 3, 2012

Response by poster: Thanks to people who posted papers with results from their field! I can always find good examples of statistics but know less about what is important in a given field.
posted by benthegirl at 6:21 PM on June 4, 2012