The New England Journal of Medicine and the Robert Woods Foundation collaborated to poll a random sample of physicians, and over 2000 of them returned the survey.  Of those surveyed, 62.9 percent supported a public option in conjunction with private options.  9.5 percent supported a public option only (aka single payer).

The process appears to have been fairly meticulous, however, there may be some grounds to challenge the accuracy of the results.  They were given 6000 random names of physicians.  They excluded physicians in training (limited experience with insurance), which amounted to about 900 of them.  Over 200 of them had wrong or incomplete addresses, or were dead.  Of the remaining doctors on the list 2130 returned the survey.

I’m not a statistician by any means, but the two factors I see on the face of the report which might skew the results towards the public option are as follows.

1.  The poll required a little bit of initiative in receiving the survey and then returning it.  Those who feel more strongly about the issue were probably more likely to return the survey.  It’s conceivable that the physicians who feel strongly about the issue may tend towards a public option, where as the less engaged physicians may be less likely to support it.  It’s also conceivable that they would support it, or be indifferent (hence their lack of response).  But I question the practice of relying on the physicians to return the survey on their own as guaranteeing a sufficiently random sample.

2.  This sentence confuses me:  “Women made up a smaller proportion of respondents than of nonrespondents (26.8% vs. 31.2%, P<0.001).”  Were they a smaller group proportionate to their numbers in the sample?  Proportionate to their numbers in the profession?  Certainly they were smaller overall, because there are fewer women than men in the profession.  But were they proportionately higher than representative or lower?  As women tend to be slightly more liberal than men in the general population, inaccurate proportionality could skew the results one way or another.

However, these factors are probably considered in the margin of error, and in any case these results are quite remarkable.  Many more details through the links, including surveyed responses to specific proposals, including the Senate Finance Committee’s 55 to 64 year olds’ Medicare buy-in option.