to spot a crappy poll // (Steven Hayward).
Podcasts and other media often feature my comments on some of the methodological issues in current opinion polling. But rather than getting into the details of sampling difficulties, statistical weighting, and question design, I’ll suggest a simpler way to instantly detect a crappy poll: check if it’s from NBC News.
The latest NBC News poll just out finds that 45 percent of voters approve of Joe Biden’s performance in office, while 47% of registered voters say they prefer Democrats to control Congress, while 46% want a Republican-controlled Congress. These numbers are closer to other polls that show much larger Republican margins as the domain moves from registered voter to likely voter, but the general conditions this year and some of the “moods” numbers in this poll make it difficult to determine “likely” voters.
The real indicator that this poll is a joke is the fact that the poll reports that the most important issue for midterm voters is “Threats against democracy” (20 percent), which is behind inflation (18 percent). This answer is not actually volunteered by voters. It is clearly pre-calculated and designed to tickle liberal voters who answer the poll.
The poll also found that Americans are deeply polarized. A large majority of Americans distrust the other party and consider them a threat to America. Gee–what profundity. These percentage numbers, along with similar findings, show that we are not in the presence any ordinary pollster.
You can read the entire 24-page poll results if you’re a glutton to punishment. It has 29 questions plus a panel for statistical purposes. However, many of the questions have six- or seven parts. This means that a respondent may only answer 75 questions. Question 16 is my favorite, asking respondents to rank their trust with several public offices, including their state legislatures. I would bet that less than 10% of respondents can name their state senators or lower house members.
A poll that is so long is worse than useless.
