11.07.2015 Views

2DkcTXceO

2DkcTXceO

2DkcTXceO

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

576 Personal reflections on the COPSS Presidents’ Awardare working on them, which seems a counter-productive strategy to me. I oncespent a fruitless year trying to define something “major,” and all I ended upwith was feeling stupid and playing golf and going fishing.I now just float: folks come to me to talk about their problems, and I tryto solve theirs and see if there is a statistics paper in it.What I do like though is personal paradigm shifts when researchers wanderinto my office with a “simple” problem. This happened to me on a sabbaticalin 1981–82 at the National Heart, Lung and Blood Institute in Bethesda,Maryland. I was a visitor and all the regular statisticians had gone off for aretreat, and one day in walks Rob Abbott, one of the world’s great cardiovascularepidemiologists (with a PhD in statistics, so he speaks our language).He asked “are you a statistician,” I admitted it (I never do at a party), and hewanted to talk with someone about a review he had gotten on a paper aboutcoronary heart disease (CHD) and systolic blood pressure (SPB). If you go toa doctor’s office and keep track of your measured SBP, you will be appalledabout the variability of it. My SBP has ranged from 150 to 90 in the past threeyears, as an example. A referee had asked “what is the effect of measurementerror in SBP on your estimate of relative risk of CHD?” In the language ofcurrent National Football League beer commercials, I said “I love you guy.”I will quote Larry Shepp, who “discovered” a formula that had been discoveredmany times before, and who said “yes, but when I discovered it, it stayeddiscovered!” You can find this on the greatest source of statistics information,Wikipedia.Iwasconvincedatthetime(Ihavesincefoundoutthisisnotexactlytrue)that there was no literature on nonlinear models with measurement error. So,I dived in and have worked on this now for many years. The resulting paper(Carroll et al., 1984), a very simple paper, has a fair number of citations, andmany papers after this one have more. How many times in one’s life does astranger wander in and say “I have a problem,” and you jump at it?Actually, to me, this happens a lot, although not nearly with the same consequences.In the late 1990s, I was at a reception for a toxicological researchcenter at Texas A&M, and feeling mighty out of place, since all the lab scientistsknew one another and were doing what they do. I saw a now long-termcolleague in Nutrition, Nancy Turner, seeming similarly out of place. I wanderedover, asked her what she did, and she introduced me to the world ofmolecular biology in nutrition. She drew a simple little graph of what statisticiansnow call “hierarchical functional data,” and we have now written manypapers together (six in statistics journals), including a series of papers onfunctional data analysis (Morris et al., 2001; Morris and Carroll, 2006).

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!