Imagine a world in which Donald Trump hadn’t won the Republican nomination in 2016. Instead, it was Republican X, a GOP candidate who falls somewhere along the political/ideological spectrum between Trump and, say, former Florida governor Jeb Bush. I don’t know what that means, precisely, but you get the point: Not establishment and not Trump, but somewhere in between.

Imagine further that now-President X enjoyed the same economy over the past two years as has Trump. Same jobs numbers, same market growth. Data from Gallup released Wednesday shows what that might mean for X’s Republican Party in next week’s midterms. There’s a correlation between views of the economy and the fate of the president’s party, anchored by the Democratic success in 1998 and the party’s poor showing in 2010.

That figure — those saying the economy is excellent or good minus those saying it’s poor — stands at plus-43 at the moment. That’s where President X would be.

What does that mean in terms of House results? Well, a rough estimate based on the very limited number of data points available suggests a slight loss in seats for the party, but not enough to lose control of the House.

This is far from certain, of course, but looking solely at the economy, past data suggest that the election wouldn’t be a wipeout for the Republicans.

The president is not President X. The president is Donald Trump — and his unpopularity is a problem.

Let’s do the same graphs with presidential approval. Here’s what the pattern has looked like, and where Trump stands in 2018.

The only president in Gallup’s polling with lower approval pre-midterm was George W. Bush in 2006, reeling from the unpopular Iraq War. Where would this indicator alone put the House results?

A loss of 39 House seats. Which, as of writing, is the average seat gain Democrats are expected to see, according to FiveThirtyEight’s forecast.

Trump makes a big difference.

There’s a sort of middle-ground metric included in Gallup’s analysis: satisfaction with how things are going in the country. This is a metric that would seem to include both the strong economy and the unpopular president.

As with the other metrics, there’s a correlation between satisfaction and how well the president’s party does. It looks like this, with 2018′s relatively low satisfaction indicated.

Why so low? Clearly the health of the economy is not the only, or even the primary, way in which Americans view the state of the United States.

Past data on satisfaction would suggest that the Republicans would lose about 26 seats Tuesday; Democrats need 23 to gain control of the House.

These are very rough estimates. You can see, for example, that the satisfaction figure in 2006 of 35 percent led to a loss of 30 seats, while the satisfaction figure of 32 percent in 1990 lost the Republicans only eight seats.

But the distinction between how the election might go were the economy the only metric and how it might go looking only at Trump is significant.

In 1998, Bill Clinton was popular, and views of the economy were good. The Democrats picked up five seats. In 2002, views of George W. Bush were good, and views of the economy were mediocre; Republicans picked up six seats. In 2010, views of Barack Obama were weak, and the economy was viewed poorly. Democrats lost 63 seats.

The closest analogue to this year — an unpopular president and decent economy — seems to be that second Bush midterm where the party lost 30 seats. According to FiveThirtyEight’s modeling, that would be a better-than-average result for Trump.