Powerline Blog: Destroying Careers For Partisan Gain I've corresponded with the Star-Tribune pollster Ron Daves in the past while taking a masters-level statistics course. I inquired about the methodology of the Star-Tribune poll, and Daves was gracious enough to answer all of my questions and provide me with a greater understanding of the polling process. I was impressed with his professionalism and devotion to his career. Because of this experience, I was angered to see him dragged through the mud for political gain. Scott W. Johnson, a TCF vice president and attorney, has been part of a concerted effort on the part of Bush partisans to silence the Minnesota Poll/have the pollster fired. Johnson and his cohorts claim that the poll unfairly favours Democrats, and cite past election results to support their claims. That's fair enough. What angers me is that they want the pollster to resign, or at the very least, to have the poll cancelled until after the election. Predictably, they prefer an ad hominem attack rather than debating the merits of different types of polling. Johnson writes, "In his conversations with me Daves has generally referred to the accepted social science research that supports the Minnesota poll methodology. Even if the poll's theoretical underpinnings are sound, however, the evidence of actual election results strongly suggests that the poll errs in practice, consistently, in favor of the Democrats by about five to seven points. It is past time that the Star Tribune is called to account". Could there be any bias on the part of the Star Tribune? Johnson is perfectly entitled to call this into question. Does he have any compelling evidence as such? Not according to any statistician I've spoken to. In any event, what more can he expect than transparency and an explanation? A true scientist would never change methodolgy due to political pressure. I don't object to criticizing Daves or the Minnesota poll, but attempting to shut it down or get Daves fired is beyond the pale. Those who believe in science would look at all the different types of polls conducted in past political campaigns, perform a comprehensive meta-analysis, and determine which approach is the most accurate based upon historical results. Johnson doesn't believe in science, or at the very least, doesn't believe it should be the primary standard for evaluating polls. He wants results favorable to republicans, and he wants them now, whether they're scientifically valid or not. This objective is worth destroying a reputation and career. The end justifies the means. Johnson's own correspondence with Daves gives him away.