Stats to ponder

I’m lucky enough to be flicking through the latest Review of Economics and Statistics so I thought I might sum up a couple of interesting tidbits while I’m reading:

  • Ted Joyce continues the battle with Donoghue and Levitt over whether more abortions lead to less crime. He crunches some (more) numbers and finds that no, they’re wrong. Again. Stay tuned for the next exciting instalment!
  • Martin Weitzman says that climate change, if it’s really, really bad, can’t be modelled using normal tests because the [insert stats talk here] and so our discounting of future costs may be mistaken.
  • University of Toronto researchers find that having a good college professor makes little difference to your grade. In fact, it doesn’t even really matter if you go to a research-based or teaching-based school!
  • Using [an advanced statistical method], Kate Antonovics and Brian Knight separate out racial profiling from racial discrimination and find that… police racially discriminate. Not exactly a stop-the-press moment, but it’s always nice to have some solid stats to back up your prejudicesviews.

And that’s your daily serving of stats talk (minus the boring numbers)!