Chances of Rain Are...
Probability statements were first included in a public forecast in 1954, and have been used widely since 1965. Given the last 40-50 years, it is expected that the public now understands such forecasts....one would think!
For example, how do people interpret the forecast prediction of a 70% chance of rain today in the Seattle area?
Multiple authors have explored this lack of understanding. Kanold suggests that the public interprets this prediction to mean it will rain. Thus, if it does not rain, the weather forecaster was in error.
- Some think that during the 24-hour period, it will rain 70% of that time, or almost 19 hours
- Others think that given a map of the Seattle area, 70% of the land area will receive rain
- And others think the correct idea...which is?
Adding more specifics, Murphy & Winkler state that the forecast lists the probability that at least 0.01 inches of rain will fall on the official rain gauge within the given period of time. And, the forecasts are quite reliable. That is, on days with a probability of 70% of precipitation, the relative frequency of precipitation tends to be very close to 70%.
So, will anything improve the public's understanding? Not as long as we tend to emphasize algebraic manipulations while decreasing time spent of probability and statistics notions. Long live misunderstood forecasts and state lotteries!
Source: C. Kanold (J of Stat Ed, 1995) and A. Murphy/R. Winkler (J of Amer Stat Assoc, 1984)