Wednesday, September 17, 2014

Walkabout Wednesday

These are things that caught my eye from the internet over the course of the last week. I used to just 'like' or share these on Facebook.

Short term projections from a model that is known to be false in the long term can be useful. In some cases, however, they can be downright alarming, as in this projection of deaths from Ebola in West Africa that accounts for the observed exponential trend. I always wondered how the WHO came up with the number of 20,000 deaths. That now seems wildly optimistic. And here's a link to a more epidemiologically serious approach; "less alarming", but still makes 20,000 dead seem wildly optimistic. Along the same lines, the word "sustainable" appears unsustainable.

Bryan McGill has another post on Statistical Machismo over at Dynamic Ecology. He's responding to the recent response to an article that's now old hat criticizing the estimation of detection probabilities. He digs into one of the appendices and makes some good points. For my part, I'd just like to point out that you CAN optimize a sampling design for something other than bias or MSE of the estimator, and you do get a different answer. Maybe this should be a whole blog post too.

Are calls for academics to be more civil in public discourse really attempts to stifle academic freedom? (ht: John Carroll)

Marcelo Gleiser argues that scientific curiosity is really an expression of spirituality. He defines 'natural spirituality' as "... a connection with something bigger than we are, seducing our imagination, creating an urge to know, to embrace the mystery that surrounds us and the mystery that we are." I like that idea a lot at first glance.

Mountain lions do attack people, and usually end up dead as a result. I think this is actually a sign of population recovery, but maybe that'll have to be a full blog post.

Turns out I was right to blame the cabal of old white guys at the top for the lack of women in our profession. If all goes well, I'll be one of the OWG next year, so I guess I'll have a chance to fix things. As a relatively involved dad with a highly successful academic spouse, I agree that being involved in family reduced our productivity. I don't regret the choice for a minute, and I have been very fortunate that the OWG in my academic unit did not hold that against me.

Redheads are bad news, at least for Gouldian Finches. What is interesting to me about the story is that they made a prediction from a model and then tested it against empirical data. (ht: Patrick O'Connor).

NPR's Morning Edition has a series on the impacts of cuts to biomedical research funding. This morning's show made a couple of good points about the perverse incentives in research that drive scientists to a) overstate their results, and b) not attempt to replicate previously published work. I think these same issues are about a thousand times worse in non-biomedical research like Ecology.

Plastic. It's everywhere in the ocean's and Britta Hardesty and Chris Wilcox from CSIRO have just finshed up the largest sampling effort ever. It's not a pretty picture

No comments:

Post a Comment