Today’s your last chance to take part in New America’s Weekly Wonk 2014 Forecasting Contest. Most of the questions are brilliant, but my favorite is #4:
Which number will be highest in 2014?
- Gold medals won by China in the Sochi Winter Olympics
- Oscars won by American Hustle
- Days the temperature exceeds 100 degrees Fahrenheit in Washington, D.C.
- Mentions of the word “progress” by Barack Obama at his State of the Union Address
I’ll pick my answers after finishing this blog post… (via Tobias Bunde)
A far more serious forecasting exercise was just published by Jay Ulfelder. He and Ben Valentino used a survey to plot the likelihood of “state-led mass killings” around the world:
It is very important to understand that the scores being mapped and plotted here are not probabilities of mass-killing onset. Instead, they are model-based estimates of the probability that the country in question is at greater risk than any other country chosen at random. In other words, these scores tell us which countries our crowd thinks we should worry about more, not how likely our crowd thinks a mass-killing onset is.
Also note that less than 150 people took part in the survey. Still, the method is neat.
In another post, Jay sheds some light on how and when he is going to post his 2014 forecasts for coups. Again, it turns out that attaching probabilities to very rare events is extremely tough. The post also illustrates that choosing a data source can be driven more by administrative reasons (who publishes when, how often, and how transparently) than by trust in its quality (or: optimal fit for the purpose)…
At the Duck of Minerva, Brian J. Phillips explores how to rank IR journals:
What are the best International Relations journals? How do we know if one journal is better than another? And how should this affect your decision about where to send a manuscript?
Phillips considers surveys (TRIP) and citation indexes (Thomson-Reuters, Google, etc.) that might be relevant to come up with a ranking. You won’t be surprised by the top 10. But the data, the discussion and follow-up questions, and the (critical) comments are well worth reading…