I think you are correct in the power of each type of setup. I think the answer depends upon the current structure and how well it's doing.
Big Site: if you have a big site and it's doing well there's no reason to break it up unless you're finding it hard to work with (SEO, navigation, or conversion).
Small network of sites: if they're doing well why combine them unless there isn't enough content/depth to them or they haven't earned enough link pop.
I believe there is a happy medium - a midway point at which the value of one or the other type diminishes as you approach larger or smaller. You could, of course, view it as a continuous process of dividing a site up. You start with one, build it until it's either unwieldy or too broad and then divide it along some logical and leveraged SEO value.
Just my thoughts but the true answer lies with the details of your site/circumstances. If you know your competitors well enough to know what sites they have, then I think I can assume you know your market space well enough to know if the products are low margin/everyday items or higher margin specialty items.
IMHO it's bad to combine many different high margin items under one store because it brings down the perceived value - the association is that the store is more of a "general store". That's my perception and it lowers my sense of the value/price of the products they sell - I expect them to have lower prices! Whereas I perceive a specialty store, smaller and more focused product line, to have higher value and knowledge. Or at least the potential to project this image is there - they could still blow it by looking unprofessional.