Skip to main content
ABC News
Why McConnell Has So Little Leverage Over Moore In Alabama

National Republicans have called on Alabama Senate nominee Roy Moore to step aside after allegations of child molestation. On Monday, Senate Majority Leader Mitch McConnell even upped his call for Moore to quit the race, ditching the if-the-allegations-are-true qualifier and saying he may also support a write-in campaign.

Moore doesn’t seem to be going anywhere. And he doesn’t seem particularly worried about McConnell’s opposition. Do McConnell and the national GOP establishment really have so little influence in Alabama?

At first glance: Absolutely. McConnell is unpopular among Republicans nationally and probably more so in Alabama. At almost any other time in modern U.S. political history, a congressional leader would likely have had more leverage with his or her party’s voters than McConnell has.

Case in point: I tried to find the worst net approval or net favorable ratings1 in live-interview polls2 for every speaker of the House and Senate majority leader since 1981.3 I then looked only at how members of their own party viewed these leaders. In the table, I’ve taken an average of McConnell’s ratings in the latest survey across all pollsters (i.e. not just live interviews) over the past three months and McConnell’s ratings among Trump voters in an Alabama survey taken this past week.

Trump voters in Alabama really dislike Mitch McConnell

How approval of Mitch McConnell among Republicans nationally and Trump voters in Alabama compares with previous congressional leaders’ lowest net approval rating among members of their own party

Mitch McConnell’s current approval rating
National (all Republicans) Aug. 14 – Nov. 14, 2017 30% 41% -12
Alabama (Trump voters) November 2017 18 66 -48
Previous congressional leaders’ lowest net approval among their own party
Tip O’Neill May 1985 61% 18% +43
Bill Frist October 2006 50 13 +37
Howard Baker July 1985 45 11 +34
Tom Foley October 1994 42 14 +28
Bob Dole February 1996 41 20 +21
Nancy Pelosi October 2010 53 35 +18
Paul Ryan May 2016 33 18 +15
Newt Gingrich October 1998 38 25 +13
Thomas Daschle January 2002 17 5 +12
Dennis Hastert October 2006 16 7 +9
Trent Lott January 1999 14 8 +6
Harry Reid October 2014 25 25 0
George Mitchell April 1992 13 14 -1
John Boehner August 2015 37 42 -5
Jim Wright May 1989 20 43 -23

All polls taken when leaders were holding their position as leader. All polls with the exception of McConnell’s were conducted via live interview. Numbers may not add up because of rounding.

Sources: Roper Center, Change Research

McConnell’s -12 percentage point net favorable rating among his party nationally is the second worst since 1981. Only Jim Wright, who had to resign from Congress because of ethics problems, scored worse, at a -23 percentage point net favorable rating with Democrats in a Gallup survey conducted in May 1989, just before he left Congress. Former House Speaker John Boehner, who felt tremendous pressure from conservatives to resign , had a -5 percentage point net favorability rating among Republicans in a Gallup survey just before he stepped down. Newt Gingrich actually had a positive net favorable rating among Republicans, at +13, in a CBS News poll before he was forced out.

In Alabama, McConnell looks like he’s in even worse shape. A survey by Change Research, an internet pollster,4 from late last week5 asked voters to rate McConnell’s job as majority leader on a scale from 1 to 10, with 1 being very poor and 10 being excellent. Just 12 percent of all voters gave McConnell a positive score (6 or above). On the other hand, 76 percent of voters disapproved of McConnell (4 or below). That translates to a net approval score of -64 points.

McConnell’s numbers are nearly as bad among Moore backers. Only 16 percent approve of McConnell’s job performance in the Change Research survey; 71 percent disapprove. He gets an 18 percent to 66 percent split among Trump voters for a net approval score of -48 percentage points. I suspect those numbers may somewhat exaggerate how many Republicans in Alabama dislike McConnell, but they mostly fit with other data we have. A JMC Analytics survey taken during the Republican Senate primary runoff found that a minuscule 10 percent of Republican voters in Alabama said McConnell’s support made them more likely to vote for Sen. Luther Strange, whom McConnell had endorsed. A near majority, 45 percent, said McConnell’s support made them less likely to vote for Strange. Remember: These are Republicans.

So Alabama voters, even Republicans, don’t like McConnell. He could, however, still affect the race. The general election polls are so close at this point that even the smallest movement away from Moore could make a decisive difference. In five surveys conducted since the allegations became public, Moore is averaging 46 percent to Jones’s 44 percent. If just 5 percent of Moore’s current backers — that would be about one-third of the Moore backers who approve of McConnell — switch their vote to Jones, the race flips to Jones being ahead 46 to 44 percent.

Then again, it’s possible that many Moore supporters will only become more entrenched in Moore’s camp given their dislike of McConnell.

This whole situation with McConnell is unusual. Moore might join the Senate despite an explicit condemnation from his party’s majority leader, perhaps with McConnell backing a write-in Republican. And if Moore wins, it seems like national Republicans may be willing to team up with Democrats to expel Moore from the Senate.

The fact that national Republicans led by McConnell have to contemplate such drastic measures stems directly from how little influence they have with their party’s voters — as documented by the fact that Moore is the nominee in the first place.

Dhrumil Mehta contributed research.


  1. These net numbers are the percentage of respondents who approve (or rate the person favorably) minus the percentage who disapprove (or rate the person unfavorably). I use net ratings instead of raw approval or raw favorable because different pollsters ask these questions in slightly different ways, which can lead to more or fewer undecided respondents. Using net helps to lessen the impact of question design.

  2. The Roper Center archive is almost entirely live interview polls.

  3. Note that some leaders didn’t have polls taken of their approval or favorable ratings.

  4. The Pew Research Center has documented that people tend to give political figures higher ratings in phone interviews than in surveys conducted online. It may be a type of social desirability bias.

  5. I’m focusing on this poll because of its timing; it’s the only poll we have on McConnell’s popularity in Alabama that came out after Moore was accused of child molestation on Nov. 9 but before McConnell said Moore should exit the race on Nov. 13.

Harry Enten was a senior political writer and analyst for FiveThirtyEight.