It might be tempting to interpret the YouGov poll's 5% margin of error — which means the real number for each candidate has a 95% probability of falling within plus-or-minus 5 points of the poll result — as an indicator that the race could theoretically be tied (given that the nine-point gap is inside the ten-point doubling of the margin of error). Tempting, but technically incorrect.
For any given polling result between two candidates you could double the margin of error to get a rough approximation of the margin of error of the difference, but an accurate estimation requires a calculation that takes into account the actual polling numbers involved. Doing that calculation on the new YouGov result using this worksheet reveals a margin of error for the difference between them of +/- 8.6%. So Romney's 9-point lead is beyond the margin of error.
A different calculation, arguably of greater interest, estimates the probability that the polling difference between two candidates is greater than zero — an actual difference. In the case of Romney's lead in the YouGov poll, that probability comes in at over 97%. He's winning Tennessee. Big shocker.
An interesting thing about the YouGov poll is that it is a two-wave survey. This latest result (pdf) is based on recontacting and resurveying the same humans in the poll's online panel who were surveyed in the first wave (pdf) a month earlier (with a respectable second wave response rate of 70%). So we get a snapshot of movement in the Tennessee electorate since roughly Labor Day. Some things we learn:
- Romney's lead among likely voters has expanded slightly from 8 to 9 points.
- A 3-point Obama lead among women in September has become a 6-point Romney lead in October.
- Among Hispanics, a 46-21 Obama lead in September (with 34% undecided) has given way in October to a 52-48 Romney lead. For that breakdown and the gender one above, the September numbers are registered voters rather than likely voters, but even so the possibility that previously undecided Hispanic voters in Tennessee have broken big-time for Romney is interesting, given how utterly dominant Obama's numbers are with Hispanic voters nationally.
- In a regional breakdown, Romney leads Obama in the "Nashville Area" by a hefty 53-36 margin. YouGov doesn't define the term or the region, but it clearly must extend well into the surrounding counties beyond blue-leaning Davidson to yield a margin that is so much bigger than the statewide result.
- 86% of likely voters say their mind is made up, which is up from 80% a month earlier.
An interesting tidbit is found in age breakouts, showing that Obama leads Romney 55-38 among likely voters aged 30-44. So there's hope for the future? If so it will only be temporary: Romney leads Obama 58-37 among voters aged 18-29. So much for the future.
The YouGov survey also polled the Senate race, finding that Bob Corker leads Mark Clayton 48-27, with 25% not sure. These numbers are almost unchanged since September. Among those who self-identify as Democrats, Clayton leads 59-10, indicating that a lot of Democrats perhaps haven't quite figured this race out or don't really care who the party's nominee is.