Michigan: Governor and Senate Races Split by Pollster

Michigan is producing the kind of polling divergence that tests how much weight any single survey should carry. In the governor's race, a New York Times/Siena University poll completed September 22 puts Democrat Jocelyn Benson ahead of Republican John James by five points, 48% to 43%, among 605 likely voters. A Torchlight Strategies survey completed just three days earlier, on September 19, shows something closer to a coin flip — Benson at 42%, James at 41%, in a sample of 643. That is a ten-point swing in the margin between two polls whose fieldwork barely overlaps, and it is difficult to attribute entirely to normal sampling noise given the sample sizes involved.

The same split shows up in the Senate race. Torchlight's survey has Republican Mike Rogers narrowly ahead of Democrat Abdul El-Sayed, 44% to 43%, while a Big Data Poll completed September 24 shows El-Sayed leading 47% to 42%. That is not just a different margin — it is a different leader, which is a higher bar than the governor's race clears. Torchlight's numbers in both contests run several points more favorable to the Republican than the other surveys in the field, a pattern worth watching if it persists across additional releases rather than treating any one poll as decisive.

PollingSource currently rates the Senate race a toss-up, a designation both surveys are broadly consistent with even as they disagree on who is ahead. The governor's race carries a Likely D rating, but the size of the gap between the Times/Siena and Torchlight numbers argues for treating that label with some caution until more polling clarifies whether one survey is an outlier. Michigan's dual-track polling this cycle underscores a broader point for readers scanning statewide races: methodology differences between pollsters can matter as much as the underlying political environment, and neither race here is settled enough to stop watching closely.