At the extremes that’s going to be true. Elections where one candidate is viewed as being unbeatable tend to have lower turnout, while turnout is highest when races are highly competitive. Keep in mind though that there’s sort of a chicken and egg situation there. High levels of enthusiasm and therefore turnout also tend to make races more competitive, while safe seats are often the result of gerrymandering and other forms of political dysfunction that suppress voter turnout and political engagement in general.

But regardless of whether a race is competitive or virtually uncontested, you typically want to be the one that’s in the lead. People are more likely to support politicians they view as popular, for much the same reason that they are more likely to pay money for movies and books that can claim to be #1 somehow. The bandwagon effect can be a powerful thing.

And voter turnout is heavily dependent on enthusiasm, and it’s a lot easier to be enthusiastic when your side doesn’t look like it’s losing. And that’s just the voting, there’s also donations, volunteers, media coverage, etc. all of which can be significantly affected by how people perceive the race to be going.

The idea that good poll numbers are somehow bad for that candidate’s voter turnout is largely a product of the 2016 election. Right up to election day the polls almost all said Hillary would win. And then Trump won. There’s a number of reasons for that, but one of the ones that gets mentioned a lot is that Democrats stayed home because the polls all said they wouldn’t need to vote.

The context is very important though. Hillary was deeply unpopular even before the DNC helped her snatch the nomination away from Bernie Sanders. A lot of people would have preferred that Trump lose, but really, really didn’t want to vote for Hillary. Those people might have been more likely to hold their nose and vote for her if they weren’t under the impression that she was practically guaranteed to win.