“If men were angels, no government would be necessary. If angels were to govern men, neither external nor internal controls on government would be necessary,” wrote James Madison in the Federalist Paper #51.

Lacking angels, Madison asserted that elections were one of the U.S. Constitution’s checks on political power. “A dependence on the people is, no doubt, the primary control on the government,” he wrote.

In midterm elections, historically, the people have followed through on Madison’s expectations. Since 1950, the president’s party has lost an average of 24 seats in midterm elections. That’s one more than the 23 that Democrats, now in the minority, need to win majority control of the House in 2018.

Forecasts from the nation’s most prominent election analysts suggest that historical patterns will likely hold true in this year’s House elections. Republicans are bound to lose seats. But how many seats? And will the number be enough for Democrats to gain a House majority?

According to one view frequently reported by journalists and liberal commentators through September, large numbers of Republicans will be swept away by a blue wave, propelled by a resurgent mass of Democratic voters eager to check President Trump.

Other news accounts and commentary from conservatives have countered that the elections will end with a whimper; the wave will be averted by a strong economy and late-breaking campaign developments that inspire Republicans voters.

How do we make sense of those two possibilities? As a political scientist, I draw from theories of congressional elections, models that forecast outcomes and expert analyses of current electoral trends. Reviewing these sources, I believe the odds favor a strong year for Democrats, but the extent of their gains is still in doubt.

Evidence for a wave

Let’s begin with the wave, or at least a very decisive Democratic victory.

Statistician Nate Silver recently estimated that Democrats have an 84.5 percent chance of winning the majority and are on track to win 39 seats. Political analyst Charlie Cook’s latest analysis predicts a gain of 30 to 35 seats for Democrats. Another summary of four different studies by political scientists reveals that Democrats are likely to gain between 27 and 44 seats. In all cases, that’s enough for Democrats to regain majority control of the House.

Those predictions are consistent with recent “wave” elections. Waves can occur when one party controls the White House and majorities in both the House and Senate, as Republicans do now. Since 1900, the president’s party has lost seats in the House in all but three of 29 midterm elections where one party controlled all three institutions.

In the midterms following presidential elections of 1992, 2004 and 2008, the party in control of government suffered well-above-average losses. For example, in 2010, after Barack Obama’s historic victory, the Democrats lost a whopping 63 seats in the House.

Most election theories assume that the number of seats lost by the president’s party will depend on the number of vulnerable seats the party holds prior to the election. Put simply, the party loses more seats when it has more seats to lose. Such a model estimates that Republicans will lose 44 seats in the House in 2018.

A second approach builds on that model by including a measure of public opinion. The more voters prefer candidates of the party opposite the president, the more seats the president’s party will lose. By looking at how much the public prefers a Democrat to a Republican candidate, political scientist Alan Abramowitz predicts that an average 7 percent difference for the Democrats in September opinion polls will produce a loss of 30 Republican seats.

A third model, developed by political scientists Charles Tien and Michael S. Lewis-Beck, adds in presidential approval and an indicator of economic well-being.

Tien and Lewis-Beck predict that a modest growth in disposable income and a relatively low presidential approval rating of 42 percent in June are too weak to offset the normal loss of seats a president suffers in the midterm. Given the number of seats Republicans are defending, they predict that the party will lose 44 seats in 2018.

All three of these theories depend mainly on national indicators; they treat the midterm election as a referendum on the president’s performance. All of the models cited above predict that the Republicans will lose enough seats to surrender a majority of House seats to the Democrats.

Evidence for a whimper

Yet, some analysts warn that although Democrats will gain seats, who will control the House remains uncertain. Kyle Klondik, managing editor of Sabato’s Crystal Ball, concludes that “the [Democratic] party is close to winning the majority, but they do not have it put away.”

Why would 2018 be a whimper rather than a wave? One reason is that the campaigns do not end in August or September, when analysts begin to make predictions. Late-breaking events could have a major impact.

For instance, Senator Majority Leader Mitch McConnell asserts that the Senate hearings of Judge Brett Kavanaugh fired up an otherwise sluggish base of Republican voters. If Republican voter turnout increases, Republicans will lose fewer seats.

Abramowitz’s study predicts that as the gap in the generic ballot of voter preference for a Democrat versus a Republican gets narrower, the Democrats will gain less ground. A net 4 percent advantage for Democrats would predict a gain of 23 seats. A net 2 percent would yield 19 seats.

Moreover, at the margins of victory and defeat, it’s difficult to assess the effects of campaign tactics. In a recent column, political analyst Amy Walter pointed out that Democratic campaign messages focus on the health care, whereas Republicans have directed the attention of Republican voters to the “scary” prospect of a Democratic majority, including the election of Nancy Pelosi as speaker of the House. Who knows for sure which message will work best?

Finally, beneath the broad scope of the national campaign are local factors that voters will consider in the 30 to nearly 50 “toss-up” races.

Currently Republicans hold nearly all of the seats in those races, and some evidence suggests that voters prefer Democrats to Republicans in those races at higher levels than national polls estimate. That would make the GOP vulnerable to major losses if a big wave breaks.

Even as national forces frame the choices between a Democrat and a Republican, when it comes to elections closer to home, voters will choose between individual candidates based on nuances such as personal qualities, campaign themes and issue differences.

Whatever happens, the party that wins a majority in 2018 will likely hold onto the House by one of the smallest margins in history – a fitting result for a closely divided nation.

Daniel Palazzolo does not work for, consult, own shares in or receive funding from any company or organization that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.