It could work both ways, but here appeared not to. Perhaps in previous elections it has, for example one wonders how much better the various minority parties would do if their polling wasn't consistently at total of less than 10% or so. But here the polls seem to have had the effect of boosting turnout significantly, perhaps even by 10% or so -- which is also the margin of victory, although perhaps the extra 10% weren't all "no" voters. It's reasonable to think that the high turnout was due to a lot of no voters finally making themselves noticed, though -- the only region with less than 80% turnout was Glasgow, at 75%, which voted "Yes".
Anything that helps to widen the political debate should be welcome. I think polls are a large part of that. It helps to know what other people are thinking -- especially if it's not the same as you think.