2 ms·

The article does make one comparison: President Obama's administration censored FOIAs 18% more often than his predecessor, newly-elected-first-year-President Ob

• noobermin 11y ago
  I think in general, media are getting better with statistics, at least when their biases are threatened (think of "fact checks" that give context to statistics cited by politicians). However, one part of making the media more mathematically literate is by introducing the concept of scale. In physics and engineering, the concept of scale is hammered into us as units or a "baseline" for comparison, or explicitly (Energy scales for an interaction, for example). Without a scale, a number is almost useless. Here, 77% sounds like a large number because it is a clear majority of the total. Here, the implicit baseline is 100%, which makes it sound terrible. However, at least in my field of laser-plasma interaction, an energy conversion efficiency of an interaction of 5% is a lot compared to the typical interaction! That is, if you lose 95% of the energy you put into the system, that's much better than most experiments! The baseline here cannot be 100%, it has to be the efficiency of the typical experiment. It would help if media learn that whenever it reports a statistic or a measurement, it needs to communicate a scale, a baseline, something for comparison. There are times, would you believe it, when 100% isn't the right baseline or the number of fingers on one's hand isn't a useful scale. (1000 people are many compared to my number of close friends but small compared to the US population). Because otherwise, people can get the wrong impression by comparing a naked number to implicit, and meaningless, baselines.