HoodaThunk?

In an article by Amy Gardner in the Washington Post this morning, there’s a reference to a poll produced by current Loudoun County Chairman Scott York that purports to show Loudoun’s voters prefer him and are unhappy with the direction the Board’s going these days. (There are other items in that story I’ll be commenting on later.) So, what does the poll say?

The poll of Loudoun voters shows that if the election for board chairman were held today, York would beat Vice Chairman Bruce E. Tulloch (R-Potomac) by more than 2 to 1.

Of course county elections won’t take place until November 2007, and anything can happen in 18 months. But the poll also shows that voters are unhappy with the direction of the board on such issues as traffic and development, York said, adding that he views that as a sign that the board’s Republican majority is in trouble.

…

According to York, 53 percent of those who responded to his poll named growth and overdevelopment as the most pressing issues facing the county. Twenty-five percent named transportation and traffic.
   
Eighty-six percent said they were less likely to vote for a candidate receiving campaign contributions from the development industry. Eighty-one percent said the pace of residential growth was “too fast.” And 77 percent said lowering property taxes was a top priority.

For the poll, York gathered the opinions of 400 people to get his results. Population estimate of Loudoun County as of 2006? Slightly over 260,000. Registered voters as of the last election? Just under 150,000. And for this population, he gets 400 people. Since we don’t know if the 400 people are 1) voters, or 2) likely voters, or 3) adults, we have to assume that’s 400 out of over 260,000. There’s a term for that kind of sample: statistically insignificant.

Of course, you have to read all the way to the very last paragraph of the story to read what that means for the poll. According to the story, that means a margin of error of plus or minus 6 percentage points. In terms of statistical analysis, that kind of margin means any conclusion derived from the poll is no better than a wild guess. Any reputable polling organization would never dream of putting out results from a poll with that small a sample and that wide an error. It would be laughed at, to say the least. Its credibility would be severely damaged. Which, I might add, might explain the fact that the story mentions the poll but not the organization who performed it. Vice Chairman Bruce Tulloch put it best in his quote for this story:

“It sounds like a very self-serving poll,” Tulloch said. “I could do a sample of 400 people and get the same results, only in reverse. Who did he call? Four hundred of his friends?”

Serious public leaders do not waste their time putting together suspect results of meaningless polls and touting them as the will of their constituents. What was the Chairman not doing when he was working this little bit of fluff up for presentation? What serious business of the Board, business the county’s residents put him in place to do (albeit by a very small margin), is sitting undone or delayed because York decided to parade a poorly done poll around as a mandate? That’s a question I’ll be asking again as we approach the 2007 elections, too, Mr. York. Better have a good answer.