Crystal-Ball Gazing on the Senate Races

Data-crunchers and political pundits are coming to largely the same conclusions about the battle for control.

Wizard gazing into a crystal ball.
National Journal
Charlie Cook
Add to Briefcase
Charlie Cook
July 28, 2014, 6:15 p.m.

Every elec­tion cycle is dif­fer­ent. Each has a unique polit­ic­al en­vir­on­ment and set of cir­cum­stances, all of which keep elec­tions in­ter­est­ing. Aside from the ex­tremely com­pet­it­ive fight over the Sen­ate, the biggest change this time around is the mul­ti­tude of Sen­ate fore­casts us­ing quant­it­at­ive elec­tion mod­els of vari­ous types; these new play­ers are join­ing the game along­side the more tra­di­tion­al qual­it­at­ive — or maybe a bet­ter term is “mul­tidiscip­lin­ary” — ap­proach that has been around forever.

While polit­ic­al sci­ent­ists, oc­ca­sion­ally joined by eco­nom­ists, have modeled elec­tions for years, their find­ings were largely ig­nored out­side aca­demia. In the last dec­ade, however, stats whiz Nate Sil­ver — who shif­ted his Fiv­eThirtyEight.com site from The New York Times to an af­fil­i­ation with ES­PN — turned his sights from ana­lyz­ing base­ball to polit­ics with con­sid­er­able suc­cess, start­ing a cot­tage in­dustry of polit­ic­al mod­el­ing for pub­lic con­sump­tion. This was fol­lowed by The Wash­ing­ton Post team­ing with a group of polit­ic­al sci­ent­ists on the Mon­key Cage and The New York Times cre­at­ing its own in-house team with a fea­ture both in the pa­per and on the web­site called The Up­shot un­der the dir­ec­tion of Dav­id Leon­hardt. The mod­els gen­er­ally in­cor­por­ate past elec­tion res­ults in each state, the value of in­cum­bency, and in some cases cam­paign fun­drais­ing and pub­lic polls, all based on past his­tory, to pro­ject what may hap­pen.

Already on the scene was my good friend and com­pet­it­or, Stu Rothen­berg, and his col­league Nath­an Gonzales at The Rothen­berg Polit­ic­al Re­port. Join­ing them are our friends and com­pet­it­ors down the road in Char­lottes­ville at the Uni­versity of Vir­gin­ia, Larry Sabato and his col­league, Kyle Kondik, and their team at Sabato’s ” Crys­tal Ball,” as well as my team at The Cook Polit­ic­al Re­port. These lat­ter three ob­vi­ously look at polls, money, and past elec­tion res­ults, but — re­ly­ing upon ex­per­i­ence and judg­ment rather than num­ber-crunch­ing alone — they also more sub­ject­ively look at can­did­ate and cam­paign qual­ity and oth­er unique cir­cum­stances that might be dif­fi­cult to quanti­fy.

Rothen­berg, as well as Sabato, Kondik, and their Crys­tal Ball team, are cur­rently pro­ject­ing a GOP gain in the Sen­ate of between four and eight seats. The Times‘s Up­shot uses the same four-to-eight range, then nar­rows it to a six-seat gain as the most likely out­come. The Cook Polit­ic­al Re­port says that a GOP gain of between four and six is most prob­ably but ex­pects the fi­nal num­ber to more likely go high­er than lower. Jen­nifer Duffy, The Cook Polit­ic­al Re­port‘s seni­or ed­it­or, is put­ting Re­pub­lic­an chances of tak­ing a ma­jor­ity at 50-50. Per­son­ally, I am a bit more bullish on the GOP’s chances, put­ting it closer to a 60 per­cent chance. Sil­ver and Harry En­ten at Fiv­eThirtyEight will switch to the al­gorithm-driv­en mod­el later this sum­mer, but their Ju­ly 15 in­form­al as­sess­ment, based on ba­sic­ally the same factors that go in­to their full-blown mod­el, point to a GOP gain of 5.7 seats, which rounds up to a six-seat gain, the min­im­um gain ne­ces­sary for the GOP to achieve a ma­jor­ity this year.

As I look through the un­der­ly­ing fore­casts in in­di­vidu­al races, my hunch is that the quant­it­at­ive ap­proaches work very well in the ag­greg­ate, in terms of com­ing up with over­all fore­casts — al­though I won­der about some of the fore­cast­ing at the level of in­di­vidu­al races. One fore­caster might rate a Demo­crat’s chances a bit high­er than I would in one race, and a Re­pub­lic­an’s bet­ter than I per­ceive in an­oth­er. Av­er­age it all to­geth­er, though, and they pretty much make sense and don’t dif­fer greatly from the as­sess­ments of vet­er­an elec­tion-watch­ers or from what most party strategists for each side privately say.

The out­lier, however, is The Wash­ing­ton Post‘s Mon­key Cage fea­ture, the Elec­tion Lab. As of Monday, it rates the Re­pub­lic­ans’ chances of tak­ing the Sen­ate at 87 per­cent, and it is giv­ing a num­ber of races that the oth­er fore­casters ex­pect to be very close or at least reas­on­ably close some aw­fully large odds of go­ing Re­pub­lic­an, al­though in a couple of cases Demo­crats are rated far bet­ter off than they are by oth­ers. Mon­key Cage gives Re­pub­lic­ans a 79 per­cent chance of win­ning the Iowa open seat, an 81 per­cent chance of beat­ing Mark Pry­or in Arkan­sas, an 82 per­cent chance of de­feat­ing Mary Landrieu in Louisi­ana, and a 98 per­cent chance of win­ning the open Geor­gia seat.

Go­ing the oth­er way, the Mon­key Cage es­tim­ates Kay Hagan has a 97 per­cent chance of win­ning reelec­tion in North Car­o­lina, Jeff Merkley a 99 per­cent chance in Ore­gon, Al Franken an 89 per­cent chance in Min­nesota, Jeanne Shaheen a 99 per­cent chance in New Hamp­shire, and Mark Warner a 99 per­cent chance in Vir­gin­ia.

We’ll see what hap­pens. But as pro­fess­or Sabato is fond of say­ing, “Those who live by the crys­tal ball end up eat­ing ground glass.”

What We're Following See More »
TWO MONTHS AFTER REFUSING AT CONVENTION
Cruz to Back Trump
2 days ago
THE LATEST
WHO TO BELIEVE?
Two Polls for Clinton, One for Trump
2 days ago
THE LATEST

With three days until the first debate, the polls are coming fast and furious. The latest round:

  • An Associated Press/Gfk poll of registered voters found very few voters committed, with Clin­ton lead­ing Trump, 37% to 29%, and Gary John­son at 7%.
  • A Mc­Clatchy-Mar­ist poll gave Clin­ton a six-point edge, 45% to 39%, in a four-way bal­lot test. Johnson pulls 10% support, with Jill Stein at 4%.
  • Rasmussen, which has drawn criticism for continually showing Donald Trump doing much better than he does in other polls, is at it again. A new survey gives Trump a five-point lead, 44%-39%.
NO SURPRISE
Trump Eschewing Briefing Materials in Debate Prep
2 days ago
THE DETAILS

In contrast to Hillary Clinton's meticulous debate practice sessions, Donald Trump "is largely shun­ning tra­di­tion­al de­bate pre­par­a­tions, but has been watch­ing video of…Clin­ton’s best and worst de­bate mo­ments, look­ing for her vul­ner­ab­il­it­ies.” Trump “has paid only curs­ory at­ten­tion to brief­ing ma­ter­i­als. He has re­fused to use lecterns in mock de­bate ses­sions des­pite the ur­ging of his ad­visers. He prefers spit­balling ideas with his team rather than hon­ing them in­to crisp, two-minute an­swers.”

Source:
TRUMP NO HABLA ESPANOL
Trump Makes No Outreach to Spanish Speakers
2 days ago
WHY WE CARE

Donald Trump "is on the precipice of becoming the only major-party presidential candidate this century not to reach out to millions of American voters whose dominant, first or just preferred language is Spanish. Trump has not only failed to buy any Spanish-language television or radio ads, he so far has avoided even offering a translation of his website into Spanish, breaking with two decades of bipartisan tradition."

Source:
$1.16 MILLION
Clintons Buy the House Next Door in Chappaqua
2 days ago
WHY WE CARE

Bill and Hillary Clinton have purchased the home next door to their primary residence in tony Chappaqua, New York, for $1.16 million. "By purchasing the new home, the Clinton's now own the entire cul-de-sac at the end of the road in the leafy New York suburb. The purchase makes it easier for the United States Secret Service to protect the former president and possible future commander in chief."

Source:
×