I recently read that for low-speed, adhesion-limited acceleration (ie up to maybe somewhere in the range of 30mph) DMUs and EMUs are pretty comparable, and far superior to anything hauled by a locomotive, even an electric locomotive. When adhesion limited, the extra weight of the diesel motors is offset by the extra traction that the extra weight provides.
It is for higher-speed acceleration where the greater power-to-weight ratio of EMUs is the biggest benefit. In other words, to take advantage of the better high-speed acceleration of EMUs, station spacing must be more than some minimum. Using lighter, european-style alternative compliance rolling stock would push the transition from adhesion-limited to power-limited acceleration to a faster speed, meaning that ligher vehicles (given the same number of powered axles) cause EMUs to lose more of their advantage in comparison with DMUs.
I don't know what the threshold is in terms of spacing where the EMU advantage disappears, but it does decrease when the stations are very close together. Fairmount does now have pretty closely spaced stations by commuter rail standards (average of about 1 mile) but I suspect the distance where the EMU advantage disappears is probably a good bit shorter than that (Maybe a quarter mile perhaps?)
Given that the EMU advantage decreases with very long station spacing (since most of the time is spent cruising at speed) and very short spacing (because of the traction problem above) it would be an interesting optimization problem to calculate the station spacing where EMUs provide the greatest advantage.
Of course there are other advantages to EMUs as well in terms of noise, pollution, operation in tunnels, maintenance costs, and perhaps fleet uniformity, but it is an interesting question nonetheless.