Tuesday, April 9, 2013

Do scientists need math?

I was exactly one sentence in to this Wall Street Journal article about how you don't need math to do science when I thought "huh, I bet this guy's a biologist".  I was right.

EO Wilson is a Harvard biologist/naturalist who leads the world in the study of ants, and he wants people to know that you don't need math to be a scientist.  Now this is a good point.  From the acronym STEM to the more colloquial ways of referring to geeks, we tend to conflate being good at math with being good at science and vice versa.  For some sciences, there really is not a good reason to do this.

On the other hand, I'm not sure I loved the execution in this article.  A few things about this:

  1. What seems to annoy Wilson most is calculus requirements.  I won't quibble with him on that.  However, I think a basic understanding of statistics is critical for any scientist...otherwise how will you read/interpret nearly any paper in your field?  Statistics is often lumped in with math, so I would have liked to hear his thoughts on this.
  2. As so often happens, Wilson left the entire field of medicine out of his discussion about science.  Walk in to any group of freshmen bio majors, and you'll find a huge percentage of them are pre-med.  Many med schools require math/stats classes for admission.  That's a big reason why these kids are taking math classes to begin with.
  3. It's not until paragraph 11 that Wilson mentions that if you're bad at math, you should pretty much stay away from chemistry and physics.  So while the headline says "scientists don't need math" what he means is "some types of biologist don't need math".
  4. He estimates that only 10% of mathematical models of biological phenomena hold any water.  Given my blog posts last week, I thought that was really interesting.  
Overall, I actually ended up agreeing with Wilson quite a bit, the caveats above notwithstanding.  Making science accessible only to those who pass several bars of intelligence is a bad move.  However, I did have to wonder if he view was skewed just a bit by being at Harvard.  I mean, how many people get in to Harvard every year who are truly bad at math?  Harvard has had a grade inflation problem for years, and it's no surprise to me that people migrate out of subjects where the grading is harder/more objective to those subjects where the grading is more subjective.  I have friends who went to Harvard.  I've seen this happen.  

It may be a nice thing to tell your professor "I'm scared of math" but how many really meant "I got a C in that, but an A in my psych class?"  A sense of wonder and curiosity about the world is wonderful, a fear of challenging yourself and failure is not.  

I guess what I'm saying is, I'd like to see some data on this before I buy this explanation.  


  1. You said what I would have, and said it better.

  2. If you're a physics theorist you need a solid background in math, because that's the language. An experimentalist, if he wants to understand what he's doing, needs quite a bit of math too. But day to day work often involves nothing more dramatic than statistics and algebra. Months or even years can go by without addressing a single integral (except automatic numerical integration in the histogramming package).

    1. From my experience, engineers don't tend to use calculus very often, either.

      But that doesn't mean calculus knowledge is unnecessary. It is foundational to the skill-set.

      And statistics are still very useful.

      (Programmers usually learn a great deal of Discrete Mathematics on their way to a degree, and it's useful knowledge. But it doesn't have the pop-culture cachet that Calculus has.)

  3. Greg Cochran at West Hunter was more scathing than thee.