For the record, his book talks to me, too (but then again, every audiobook seems to do that).

Do you? (Sorry…)

NOPE.

In fact, here’s one I had originally planned on posting in response to Jake’s “Think Different” post he recently made. It made me laugh but I’m not sure how much others liked it.

http://img.skitch.com/20081014-fc9wft5ce2bixdjpw8pxd3c1up.jpg

First, here's a graph of what I think you're describing.

http://img.skitch.com/20081009-8tqfpcw54q17gnpx...

It shows that your enrollments somewhat follow the normal curve. I also realize that it would be very helpful to view the cut-off values for each of the groups. However, in my limited knowledge of all things SPSS, I can't figure out how to zoom in on the graph. My stats class, btw, is as much about learning how to use this monster of a stats program called SPSS as it is about learning stats itself.

Second, another statistic that will work well for what you're trying to show is called a Z-score. In non-propeller-head English, a Z-score tells how a single data score compares to average of all of your data. This graph was created by plotting the zscores for each of your school enrollments.

http://img.skitch.com/20081009-peaep9ty15hm38aw...

From the graph, we see that most of the enrollments are just under the average (mean) of all the enrollments. However, there are a few outliers in your sample. Those scores clear out by the 15 mean that there are some points in your data that are 15 standard deviations from the mean.

With Disqus, though, I love the ability to reply to comments via email. I love the threaded posts. And I love how it track's each user's comments, forever. Or maybe that should be “forever?”.

The only thing that I can see missing from Disqus is the ability for non-account holders to subscribe to comments via email. I've bugged them about this and they promise that it's a feature that will be coming soon.

-Sherry