It’s true, a month out from party primary elections, it’s uncertain which of three serious Republican candidates will advance to an inevitable runoff. And there’s a polling-based reason for that for which a little historical data and common sense can address partially to give a better idea who’s likely to be part of that duo.
In the last several days, the campaign for Treas. John Fleming released one, and two others have come from other organizations, although there are no details as to who paid for these. Rep. Julia Letlow was named the leader in those, while the Fleming campaign had him on top with incumbent Sen. Bill Cassidy running second. The other polls had him running third.
This continues a pattern where campaigns release polls favorable to their candidate, which may be why Cassidy’s camp hasn’t released any recently. None of the most recent trio had him at better than 26 percent, which is abysmal for a sitting senator who by himself already has spent probably at least $10 million on his reelection (as of the end of the year; the most recent numbers through the first quarter of 2026 will be available later this week) and has had surrogate political action committees as well churning out cash on his behalf.
The immediate question, obviously, is why are the results so varying, with these latest indicating Cassidy from 19-26 percent, Fleming from 23-34 percent, and Letlow 19-31 percent. As previously noted, many factors such as question wordings and ordering could play a role. However, the most prominent of these is sample composition, specifically in the mix of Republicans and unaffiliated voters that will show up.
That is not much more than barely-informed guesswork at this point. For Louisiana specifically, the only baselines that exist are 2008 and 2010, when closed primaries existed briefly, and then back to 1972 and before in an era where federal election candidates operated in an environment of extreme partisan imbalance. The latest two cycles had similar rules regarding primary participation as with this year’s, but with one critical difference.
The previous primaries under state law allowed the governing authority of recognized parties to decide whether to let unaffiliated voters participate in these. Then, Republicans chose not to while Democrats chose to let them in. (Constitutional jurisprudence doesn’t permit a state to dictate to parties that they must have purely closed primaries, or party registrants only participating.) In that sense, only Democrats’ contests provide any guide to what may come now, but even this is contaminated because of the much higher number of Democrat registrants then that change the dynamics now.
To make matters worse, 2008 data essentially is unusable to guess turnout. Because of hurricane activity, elections were pushed off their intended schedule and the Senate race didn’t feature any primaries. Only 2010 data, which did feature a competitive Senate race with primaries and an election day isolating the congressional primaries from all but a handful of very local items, is close to comparable. And even here, there was a Libertarian primary and the other two recognized parties of that era didn’t have a primary so even these numbers will be a little distorted.
And if you’re a fan of voter participation, hold on to your hernia belts. Keeping in mind that non-major party voters could vote only in the Democrats’ primary or if Libertarian the Libertarians’, Democrat turnout was 7.13 percent and “others” (unaffiliated plus Libertarians) 1.66 percent (the Libertarian primary attracted fewer than 2,500, so adjusting for that the unaffiliated turnout falls to 1.13 percent). That suggests 86 percent of primary voters in 2026 will come from party registrants.
Regardless, with a single exception, problematically none of the polling done now has revealed its GOP/unaffiliated ratio. If they report at all beyond just the candidate proportions, these (as in the case of Fleming’s latest) might report the sample size and a generic assertion that they included in it “likely Republican and unaffiliated voters.” The one exception, independently done in February showing Fleming with a healthy lead, in fact used this 86/14 ratio.
But that’s really stale now in a contest with lots of advertorial jousting since, with the precise idea of activating what would appear to be a huge latent coterie of voters to buy voted disproportionately for the favored candidate. Assurances, whether genuine and/or realistic, by those who claim the mantle of campaign consultants and who get paid regardless of outcome that this approach will change the balance enough – presumably Cassidy benefits the most from pumping in more unaffiliated voters (and may be seeking Democrats at least in the short term to declare themselves as such to vote in the GOP primary for him), followed by Letlow, with Fleming as the most consistently conservative candidate by his record in Congress (having served in the House prior to Letlow) benefitting the most from fewer unaffiliated voters – should be accepted acknowledging their hazardous nature.
Both Cassidy and Letlow and their surrogates have run very broadcast-heavy, indiscriminate campaigns. Moreover, these have trended increasingly negative in sum. The problem with this is unaffiliated voters, the vast majority of whom are so because either they are minimally interested in politics or they are but also disappointed in the major parties in often perceiving their identifiers as too prone to conflict, see these ads as either validating their disinterest because the ads largely don’t give reasons to vote for a candidate or exemplifying the very conflict that has them casting a pox on major parties.
By contrast, Fleming’s media is very focused on ideological audiences and he engages in a good deal of retail campaigning to more ideological gatherings. Keep in mind that, even as the extraordinary amount of advertising almost certainly will boost overall turnout well beyond the 12.93 percent Republican turnout in 2010 – perhaps even more than double – that subset of voters will be demonstrably unrepresentative of Republicans as a whole in that they will comprise disproportionately of ideological voters, which is precisely the demographic he tries so hard to cultivate through a strategy of mainly targeted radio and other electronic communications and retail politics. These people also are more likely to communicate to others or be solicited to communicate their opinions about the election than the voter mix the other candidates prefer.
That polling results within days of each other wildly vary most likely signals different sampling mixes (which also may interact with how partisanship is identified: by registration or by self-identification) based upon imagined campaign scenarios. Yet from the best data available, at this time the most likely electorate mix probably will produce what happens when taking the ranges in the three latest polls – Fleming leads Letlow into the runoff, and the sitting senator Cassidy starts making his post-Senate plans.
No comments:
Post a Comment