A tied poll in South Dakota narrows our margin, not the odds

An Impact Research survey of 500 likely voters in South Dakota, fielded August 13-17, put Republican incumbent Mike Rounds and independent Brian Bengs level at 44 percent each.

Our model absorbed the number and moved, but not far. Our current estimate of the margin is R+19.6, with a 5th-to-95th-percentile range running from R+30.4 to R+8.7. Our previous run, from earlier the same day, had the margin at R+23.9 with a range of R+34.4 to R+13.4. The poll pulled the expected margin about four points toward Bengs and widened the plausible low end of Rounds's win.

It did not change the bottom line. Our model still gives Rounds a 100 in 100 chance of winning the seat, and gives the Democratic column a 0 percent chance, with no Democrat on our candidate list here. That is what a single 500-respondent poll can and cannot do in our framework: it shifts the central estimate in a state where our prior is a wide Republican margin, and it leaves the outcome distribution short of anything competitive. Even the most Bengs-friendly end of our range, R+8.7, is a Republican win.

Nationally, our average of 23 generic-ballot polls over the past 45 days stands at D+6.6. In the Senate, our model gives Democrats a 45 percent chance of winning control and Republicans 55 percent, with a mean of 50.0 Democratic seats. Democrats need 51 seats for control and Republicans need 50, because the vice president is a Republican and a 50-50 tie goes to the Republicans. South Dakota is not among the seats our model treats as deciding that question, and this poll does not make it one.