Just found this. Ordinarily I’d save it for next Friday, but it’s time-sensitive so I’m putting it up now. Early Career Ecologists has come up with an intriguing way to survey ecologists’ preferred statistical methods. If you sign up to participate, you’ll be given a small dataset and tasked with explaining the variation in the response variable (acorn counts) in terms of the predictor variables (environmental variables, year, and other attributes), using whatever statistical method you want. The idea is to get a sense of the range of statistical methods ecologists use, and why.
Obviously, the range of responses probably would change a lot if a different sort of dataset were provided, or if more/less/different guidance were given as to the scientific goal of the analysis. But still, it’s a neat exercise and I’m curious to see what they come up with. I’ll probably sign up myself.
Follow the link to sign up; deadline is May 29.
This is reminiscent of, but actually quite different from, an exercise an NCEAS working group did way back in the early days of NCEAS. They made available a time series dataset, I believe on the abundances of various life history stages of some unspecified species. The contest was to build a model (purely statistical, mechanistic, whatever) that would predict the (unrevealed) future dynamics of the species. I don’t recall the outcome, and I may be misremembering details, but that was the gist. If enough folks were interested, I could see running a similar contest at some point on this blog. Might even be fodder for a group paper if we did it right.