One of the things we look for in a new site is enough interested participants to make it viable. We now ask people to make their interest known, because votes don't necessarily correlate with intended participation. Sometimes an upvote means "I want this" and sometimes it means "yeah, that sounds like a good idea (but not my area and I wouldn't participate)". Votes, by themselves, don't mean much.
What is much more important than voting is the feedback that a proposal gets -- positive, negative, and absent. We read all the answers and comments. If people have reasons for opposing a proposal and express them, we consider that. If they just downvote, that's much weaker signal -- same as with upvotes. A controversial proposal will lead us to tread more carefully, but downvotes alone aren't what blocks proposals.
In general we want communities to define their own scopes and identities. However, as hosts of a network of communities, we also pay some attention to the bounds among communities. If a new site overlaps an existing one without clear differentiation, that's going to confuse our users so we'll ask for more work to be done. For example, we have an Outdoors site; if somebody proposed an Extreme Sports site, scope to include hang-gliding and bungee-jumping and rock-climbing without a harness, we'd ask them to address the overlap. We have a Writing site; if somebody proposed a Technical Communications site, we'd likewise ask them to address that overlap. That one actually happened on SE, but SE didn't address the overlap, instead allowing a site to proceed that was 95% a subset of another one. The new site failed in private beta and lots of people were frustrated by the whole thing. We don't want to repeat that mistake.