Not possible on short notice, difficult even with a few days' notice, because it screws up the downstream route structure. Even if a 757 (for example) is available to replace an MD-80, the added cost of flying a 757 over an MD-80's planned routes - and, except for the first leg, carrying loads that would fit in an MD-80 - until they can get things straightened out again outweighs the cost of bumping a few people.

As for the 50-seater versus 200-seater question: airlines assign aircraft based on expected passenger loads, then sell tickets based on the capacity of the assigned aircraft, so the average chance of a bump is the same. However, since the law of averages tends to work better with large numbers, the statistical forecasts are more likely to be wide of the mark in either direction on the smaller plane. The distribution has the same mean, relative to aircraft size, but a wider spread. If they oversell a smaller plane, or run into some other problem that overloads it, it's more likely to be a doozy and they'd have more trouble getting enough volunteers. All in all, though, I wouldn't base bump planning decisions on aircraft size.