NNT

"The NNT" is a great example of taking very complicated scientific studies and presenting them in a compelling, easy to understand format. This group of doctors picks a particular treatment, reads all the studies, and then provides some very straightforward conclusions. The NNT (number needed to treat) is statistically how many patients have to be given a treatment to benefit. For example, steroid treatment for asthma attacks helps 1 out of every 8 patients,  while aspirin to prevent heart attacks helps 1 out of 50 patients. Then they provide a simple stoplight review: green is good (more benefit than harm) yellow is 'we can't tell yet', red is 'no benefit found'. They follow up with some quick statistics on benefit / risk, and also translate the statistics into words for the arithmophobic, then some specific details about the studies and how they drew their conclusions.

I like the startling factoids (1 in 2 women got a false positive mammogram result - definitely not a test worth taking). I think it's useful to know how ineffective treatments are (even for the best treatments, usually more than 90% of patients are not helped.) But more generally, I think it's amazing that I can't even decipher the titles of most of the studies they present, but I couldn't find a single thing on The NNT.com that I didn't understand. There are definitely some down sides to people getting medical advice from the web, but this site is reliable, straightforward and useful.

The cost of college

Almost 2 years ago, the NY Times published this scary looking graphic (left) about college fees are skyrocketing out of control (taken from this report), along with an article about how most people won't be able to afford college soon.

So I was kind of impressed playing around the the Chronicle's interactive graphic  (right) published a couple of weeks ago. It's fun to see how the cost of one school have changed over 12 years and how they compare to other schools. But what really stood out for me is that the interactive data really doesn't look scary at all. Sure, you can find one school here and there that had a cost spike in one particular year, but there's no pattern of precipitous climbs or exponential rises.


Why is that? Both sources are supposedly comparing the same data, the total cost of college (tuition and fees, not counting any financial aid) and both are looking at a fairly comprehensive sample of schools. So why does one generate attention-grabbing national headlines while  the other is read only by specialists? We often tend to gloss over the idea that data aren't really impartial facts. They are evidence we scientists (economics is the "dismal science") use to tell a story. This is a good example of how you can use the same evidence to support various conclusions. Even when you're not overtly biased, you have to bring a personal interpretation in going from raw data collection to results. In this case, the first study took the increase from  1982 to 1984 and calculated the % change relative to this. The second study shows the % change between only adjacent years (or the actual costs or a range of years). The change from 2009 to 2010 is smaller than from 1982 to 2002, which isn't at all surprising.