Tuesday, October 16, 2007

Judging the Autism Charts Challenge

When writing this entry on the difficulty of displaying data when diagnosing Autism at age 2, I did not expect such a large number of people putting their minds into making the graphs better. Andrew Gelman and Masanao Yajima made a challenge out of it. It got picked up by the folks at JunkCharts, EagerEyes.org, and at Perceptual edge and much brain computing cycle went into this Challenge. Previously, I had already stated that I thought Autism was a Grand Challenge and this is in no small part because the diagnosis phase is not early enough, not accurate enough and sometimes not useful enough.
  • Not early enough is the reason why this study [1] by Catherine Lord was performed. Instead of taking a "normal" test at age 3, age 2 is preferred with regards to intervention.
  • Not accurate enough and not useful: there is currently very little difference whatsoever in treatment between kids that have been diagnosed Autistic or with Pervasive Developmental Disorder- Not Other Specified (PDD-NOS).
Eventually, one cannot even devise clinical trials for drug development until some of these tests are better refined. The figures entered as part of the Challenge can be seen in full scale here:

Please go to Andrew Gelman and Masanao Yajima's blog and answer the survey at the very end. Please note there are TWO questions.
  • Which plot did you like best?
  • Which plots did you find useful?
Please answer them both.. Thank you to Antony Unwin, Stack Lee, Robert Kosara, Patrick Murphy, Andrew Gelman and Masanao Yajima for contributing to this effort.

[1] Autism From 2 to 9 Years of Age, Catherine Lord, Susan Risi, Pamela S. DiLavore, Cory Shulman, Audrey Thurm, Andrew Pickles. Arch Gen Psychiatry. 694-701, Vol. 63, June 2006

Liked this entry ? subscribe to the Nuit Blanche feed, there's more where that came from


Anonymous said...

Well thank you for the challenge! There is more I wanted to do with this data, but I just haven't found the time to do it yet. I'm thinking that the false positives vs. false negatives would be interesting to investigate further, because I don't feel we have really gotten to the bottom of that yet. But we have shown that there are a lot of much better alternatives to the Venn diagram ;)

ctwardy said...

You do plan to submit these back to the original paper author, right?

Igor said...


Absolutely agree on the Venn diagram usefulness.


I have mentioned the challenge to Catherine Lord but I have not heard back from her. But I intend on pressing this issue as she has gathered additional data since that paper was published.