fb pixel

Log In

Reset Password

Since You Asked: When average is not normal

Have lots of questions including whether most of the Trib staff is from bAshland (short for Bush bashers of Ashland) but really would like an answer re: your weather page. You show "normal" temperature or precipitation for this date, or month, etc. How do your weather people decide on normal? Notice that a couple of TV weather people use "average" rather than normal.

Just sign me:

— Tom from Rogue River

Well, Tom from Rogue River, we get accused plenty of being Bush coddlers as well, so we figure we must be doing something right, pardon the pun. And even if you don't include our freelancers, we are in pretty much every nook and cranny of the valley — the staff's hometowns roughly approximate the population distribution in the Rogue Valley. We didn't plan it that way, but that's how it is.

As to your other question, our weather people are your weather people. Our information comes from the National Weather Service via Accuweather, a company that provides weather pages for a lot of newspapers. And "normal" is the language the National Weather Service frequently uses when "average" isn't a fair measure.

We spoke with Mike O'Brien, a meteorologist with the National Weather Service office in Medford, who explained that "normal" is used to give a good indication of what temperature you might expect on a given day. Instead of averaging the temperatures, they plot all the temperatures on a distribution chart and the "normal" is the most frequently occurring temperature for that day. It's sort of like a median. O'Brien said averages "have their use" and are used frequently.

O'Brien pointed out that averages become problematic say, for instance, when you have a date over eight years that is 70 degrees, and in two years that date has a temperature of 100. The normal for those 10 occurrences would be 70 while the average would be 76. As you can see, averages can skew things. And as O'Brien reminds us of that famous quote, "there are three kinds of lies: Lies, damn lies and statistics."