Also this week: why academics write badly, haters gonna hate (and that’s a good sign), college enrollments are declining (and that’s a good sign), combinatorics vs. the h-index, the fishy exact test,
chickens CHICKENS, and more. Oh, and leopard+gravity vs. impala. And leopard vs. Marmite.
Scicurious had a really nice post on her love of National Geographic, which started when her grandfather gave her a subscription. She talks about moving through the years, always moving all those boxes of NatGeo. I love the solution she came up with for preserving the memories in those golden covers. We also had National Geographic growing up, and I also loved looking through them. This is the cover that stands out in my memory. Our National Geographic subscription began as a prize my oldest sister won (I think for winning a school spelling bee). My father likes to joke that it was the most expensive prize we ever got, since he renewed the subscription for many years after that gift subscription was over.
Why research fails. A brief and broad discussion of the many ways a research program can go off the rails, but still a good one for beginning grad students to read.
And another from Claudia Sahm: smart people don’t always know the right answers, but they know the right questions. The examples are from economics, but the point is broader.
Graduate students should not fall into the cult of being “too busy”. Nice post on how to “deprogram”. I particularly like the advice on knowing how to stop once some bit of work is good enough, and the advice to seize (and create) opportunities rather than passing them up because you’re “busy”. In my experience, the best graduate students (by any metric you care to name) are the ones who attend seminars, meet with visiting speakers (even those whose work is unrelated to their own), organize reading groups, take classes just because they sound interesting, etc. Here’s another nice post making the same point. And no, you don’t have to work 80 hours/week to do all that. (ht @kerryecharles)
Self-described “alpha female” Cathy O’Neil (“Mathbabe”) on how she figures she must be doing something right if people hate her.
The latest salvo in the ongoing debate over alternative rationales for conservation: Richard Conniff on how he’s tired of pretending that every species is or might be “useful”. A semi-related old post, and here are two more.
Some of you may recall a high profile PNAS paper from 2004 purporting to show that diverse teams of problem-solvers outperform less-diverse teams of experts. Yeah, not so much. A case study of how to abuse mathematics. There are plenty of good arguments for “diversity” in many contexts, but it’s a shame that this bad one has gotten a lot of play. (ht Mathbabe)
Let’s turn from that last link to a better use of math: to a good first approximation, your h-index is proportional to the square root of the number of times you’ve been cited–and you can use combinatorics to prove it. Technical in places, but well-written, and the intuition is straightforward. A good example of using math to aid interpretation of data. (ht Mathbabe)
Steven Pinker with an astute discussion of why so many academics write badly–and how they can write better.
In a new paper, philosopher of science Jonathan Birch tries to unpack the kerfuffle over kin selection prompted by Nowak et al. (2010). Very accessible. I read it with interest for the thoughtful discussion of what the Price equation can and can’t teach us about evolution. Includes interesting comments on trade-offs between different desiderata of theoretical models (e.g., between generality and predictive power, and between mechanistic explanation vs. explanation in the sense of unification). A nice example of a philosopher moving a scientific argument forward, by clarifying conceptual points that the scientists themselves had been struggling to articulate.
Nice blog post from Adam Algar discussing Helmus et al.’s recent Nature cover article. Human activities are reshaping anole biogeography in ways that are consistent with island biogeography theory. Notes that this is a nice example of fundamental research turning out to be relevant to an applied issue in ways that weren’t anticipated at the time the fundamental research was being conducted.
Key points about the Ebola outbreak.
Speaking of disease outbreaks, Ottar Bjørnstad and colleagues at Penn State are teaching a MOOC on infectious disease dynamics. It just started, apparently there’s still time to sign up.
My Calgary colleague Steve Vamosi has started blogging. His first post is on p-values and model selection.
Why declining US college enrollments are a good thing (well, a symptom of a good thing).
This week in artificial selection: today’s chicken breeds grow 4x larger on the same diet than breeds from the 1950s. One of those studies that’s best summarized by a photo rather than a figure.
As a blogger, I enjoyed this: the 10 best essays since 1950. Not because I think it’s a definitive list (I haven’t read nearly enough essays to judge), but just because I always like being pointed towards good essays. Longer blog posts are like essays, or perhaps mini-essays. (ht Marginal Revolution)
An amusing online demonstration of fishing for statistical significance. (ht Andrew Gelman, though some of his commenters on this seem to have missed the joke)
Sub-optimal foraging: a leopard tries Marmite. The reaction is…hard to interpret. 🙂 (ht Not Exactly Rocket Science)
Presumably this is the hunting technique the leopard used to catch the Marmite. Wow! 🙂 (ht Not Exactly Rocket Science)
Hoisted from the comments:
A recent linkfest led to a good discussion of the ethics of lecturing, as opposed to using active learning.
Just a quick note to say how much I liked the Birch article. After reading Nowak (2010) and some of the responses and defenses for a population genetics seminar, I felt like I simply didn’t understand the issue. This article help me move forward in understanding at least a little more. Thanks!
Always interesting to see which links are most popular. This week it’s the cult of “too busy”, and “why research fails”.