Wednesday, June 18, 2008

Cognitive dissonance with graphs

This is from the Wall Street Journal blog, who has more money for copyeditors than I do. Leaving aside my distaste for pie graphs, take a look at what's wrong with this graph.

Because this is such a major issue, you would think that someone would catch this mistake before production?

But, in the spirit of Junk Charts, I'll offer the way I think this should look. I think the insured/uninsured/underinsured lines need to be plotted over time on a single line graph to support the author's point, which is that the underinsured population has grown over time.

PS. The original report is here. All of the supporting displays are eye-gouging tables. I made the chart I suggested, and even with Excel Defaults, it beats the heck out of the pie chart.

Saturday, June 7, 2008

How safe is safe enough?

So, another short post while real life takes it out of blogging.

Is 100% safe enough? 99% safe enough? Let's make this more specific and say that the risk being talked about is irreversible morbidity or death. How safe is safe enough?

Hard as it is to believe, I'm linking to Orac a second time. Here, Orac discusses risk in the context of the Green our Vaccines rally.

I think what will be useful is a table, a direct comparison of risk of irreversible morbidity or death for those who have been vaccinated vs. who have not been vaccinated. Of course, herd immunity is going to confound this table somewhat, but I think such a table with some explanation about what can go wrong under each scenario will be very useful.

Epi wonk?

Tuesday, June 3, 2008

Do you think Bayesian?

Because I'm in the weeds right now, I'm not finishing up my third installment of blinding and randomization. I will, however, refer you to an blog entry that discusses the notion of whether our learning faculties implicitly use Bayesian statistics.