The first time I ever watched a State of the Union address was when I was in college. It was the Bill Clinton era, and I figured that, as a sort-of-adult I should tune in. I did. It was fine.

The point of that extensive anecdote is this: People tend to watch a State of the Union address for one of two reasons. Either they really like the president and want to hear him speak, or they do it like it’s homework. Which often means that the audience is weighted toward people who tend to view the president positively.

Snap polls conducted after President Trump’s joint speech to Congress — not a State of the Union, since it is Trump’s first year — suggest that he got strongly positive views. Among all respondents to a CBS poll, 59 percent said that they strongly approved of what Trump had to say. In CNN’s survey, 57 percent said they viewed it very positively. Those numbers vary by party identification, but were generally strong.

But, again, much of the audience was made up of people who like Trump — including independents and Democrats. (In CNN’s poll, for example, the pool of respondents was 8 percentage points more Republican than the population on the whole.) Other first-year speeches received similarly glowing reviews. In CNN’s poll, George W. Bush’s 2001 speech was viewed very positively by 66 percent of respondents. Barack Obama’s in 2009 was at 68 percent. Both of those numbers, you’ll note, are higher than Trump’s.

The question for Trump is less about the effect of one night’s speech and more about whether he can turn around his exceptionally low approval ratings. In Gallup’s most recent weekly average, Trump was at 42 percent — up slightly from recent weeks but still well below other new presidents.

Do these speeches to Congress actually shift that trajectory? Well, a little. Sometimes. Briefly.

Let’s start by looking at Clinton’s numbers from his eight speeches. (For Clinton and Bush, we used Gallup weekly numbers from the closest week before and after the speech.)

In 1993, Clinton saw a nice little bump in his approval, across the board. That held true in most years: The closest week before the speech was lower than the closest week after five of his eight speeches. His numbers among Republicans always improved — which is a good sign for Trump. After all, for the most part, Trump’s approval numbers are so low because he’s viewed negatively by Democrats (and unusually negatively by independents). An increase in his approval numbers among those who oppose him could be significant.

But that was Clinton, a guy well known for his oratorical skills. George W. Bush had a slightly different experience.

After only three of his speeches did Bush see improvement, and only once among Democrats. But that improvement with the opposition party came after his first speech in 2001, to the point above. (Of course, in 2002, he had nowhere to go but down, seeing massive approval ratings in the wake of the 9/11 attacks.)

Obama’s experience was remarkably mixed.

In half of his eight speeches, he saw an overall gain the week after vs. the week before. In half, Republicans improved their opinions of him. In half, Democrats did. But, again, look at those numbers among Republicans in 2009. A big jump the week of the speech — that then faded.

This pattern suggests that Trump’s year-one speech could goose his support among Democrats, at least for a little bit. Assuming that enough of them felt obligated to do their civic homework and tune in.