This morning I finished my last mid-term and I wanted to celebrate, but unfortunately rainy weather limited my options. In order to find out the likelihood of my parade being drenched, I took a look at our online weather report. It said there was a 50% chance of rain.
Then I realized that I had no idea what that statistic meant. Did it mean that there was a 50% chance that it would be raining during any particular instant of the day? Did it mean that 50% of Taipei would get a measurable amount of rain? Did it mean that there was a 50% chance that there will be rain at some point during the day?
The answer is, rat-a-tat-tat, that of previously recorded days with similar weather conditions, 50% of them had a measurable amount of rain at some point during the day.
I looks like I am not the only one confused. I pulled the above definition from a study measuring what people in various cities in Europe and the United States think “x% chance of rain tomorrow” means. The researchers gave people three choices: Time (x% of the time it will be raining), Area (there will be a measurable amount of rain in x% of the applicable area) and Days (x% of previous days with similar weather conditions had a measurable amount of rain sometime during the day). As I mentioned above, the “Days” option is the correct one.
Here is a chart reporting the author’s findings:

Go New York! Interesting how people in Amsterdam were mostly wrong in the same way, while in Athens and Milan the distribution was more even.
Full disclosure: I would have guessed “Time” before I saw the answer.