Math Stories

{ 4 } Comments

1. Mr. K | February 14, 2009 at 9:35 pm | Permalink

   To answer my own question (following some research) apparently nothing.

   What's wrong, or perhaps just inconsistent, are the California math standards. According to them, all data points must be included between the ends of the whiskers. But, and this seems to make sense, you're allowed to throw out the outliers from the dataset, and include them as separate points.

   Now I'm going to be all conflicted the next time I have to teach this.

2. Sarah | February 16, 2009 at 6:31 pm | Permalink

   Laughs at silly standards changing the way we present data. You got me thinking. I'm not convinced the scaling works. Is the outlier really twice as much as 1.5 times the 3rd quartile?

3. Druin | February 19, 2009 at 6:52 pm | Permalink

   The boxplot shown here is called a "modified boxplot". The "regular boxplot" is the one described by your standards. With a modified, you test for outliers and show them as separate points and then the the whiskers go out to the highest (or lowest) non-outlier point.

   Of course, you may not have been serious in your question and I'm just too tired to get it :) Sorry about that.

4. Mr. Follett | March 4, 2009 at 7:05 pm | Permalink

   Thanks for sharing this. I re-posted it.

Post a Comment