It's literature that I used to know better 15 or 20 years ago, so I don't want to characterize all of it, but I'll say two things about it. One is that ballot design effects are generally small. Very rarely do ballot design effects matter for the outcome of an election. I think we should be clear about that.
There are, however, a few notable irregularities. One is that ballots that are alphabetically ordered favour people who are at the top of the ballot. There's fairly well-established empirical evidence that ballot order matters, so people who are closer to the top of a ballot matter more. The length of the ballot is thought to increase those effects, because the longer a ballot is, the more likely you might be to say, “Oh, come on. This is too much. I'll just go for someone near the top.” That said, a lot of that comes out of studies in the United States, where you have a lot of different offices that you're voting for at once, for example.
Here's the most notable case. There's pretty good empirical evidence that in Palm Beach County in 2000, Al Gore lost and Pat Buchanan won that county because of the ballot design. They had a so-called butterfly ballot, which opened up in such a way that it was unclear which person you were voting for on which side of the ledger. This is a famous example of how a ballot design can lead to unintended effects.
To make a long story short, I think the effects of this would be small, but the frustration for voters could be real in having to confront a ballot with hundreds of names, which we've never really seen before in Canada.