Several of the comments here give good insight into why ditching was often preferred.
Several people have suggested the value of knowing / controlling where you'll hit the water. If you want to be close to the rescue ship, ditch. If you want to be close and you bail out, you may wind up in the radio masts of the rescue ship, or under it's bows. WW2 parachutes could not be steered. And, in addition, the plane you just bailed out of, near your friendly rescue ship, now has to go off and find it's own place to come down. Near your friendly rescue ship. Maybe too near. Maybe you miss getting tangled up with the rescue ship, but your plane doesn't. Altogether not a good solution.
Several others have also described the value of the raft. Most USN aircraft (can't say as I know about RN or IJN or …) did carry rafts. If you are going to be in the water for an unknown period of time, the raft is a HUGE advantage to survival. So yes, you want to be with the plane in the water for a few minutes, if you can.
Most USN aircraft were designed such that they gave a bit of time before they sank. Well, I shouldn't say design … because I don't know if it was in the design requirements. But it was clearly a feature that the Navy took seriously, as they chose to buy additional aircraft that displayed this characteristic, and not to acquire additional aircraft that did not. With one notable exception (see below).
But there is another reason not mentioned -- parachuting into water is downright dangerous. I'm sure any of our members who are naval aviators, or Coast Guard, or even trained parachutists will be able to correct / enhance my description. It's been almost 40 years since I did parachute training … but here's the basics of parachuting into water:
You come down into the water feet first at a bit of speed -- your downward speed with a fully deployed WW2 style parachute would typically be something in the area of 35-45 fps (25-30mph). That's why you had to "role" when you landed on the ground … if you tried to land on your feet you could easily break your legs/hips/back.
If you hit the water feet first at that speed, you're going to go about 10 or 15 feed deep before your downward movement stops. Then you have to swim back up to the surface quickly to avoid drowning. Except -- you have a great big wet silk blanket spread all over the surface above you, blocking your path to that oh-so-critical air.
It's REALLY easy to drown when you parachute into water.
The recommended solution is to "cut away" from your parachute just before you hit the water. Parachutes have quick release mechanisms for the straps, so you pop those at about 10 or 15 feet. You hit the water going about 50% faster, so you go deeper, and have a bigger job of swimming back to the surface, but the sudden release of weight from the parachute straps tends to cause the canopy to collapse to one side, making it far less likely for the 'chute to be above you blocking your path to the air.
The problem is that it is VERY hard to judge distance when you are in the terminal stages of a parachute landing. You're ability to judge distance changes dramatically between 100 feet and 5 feet (as your binocular vision comes to dominate your assessment). Most people can hardly tell the difference between 100 feet and 80 feet, but can clearly see an obvious difference between 10 feet and 5 feet. The result of this physiological phenomenon is what parachutists refer to as "ground rush" -- the ground seems to accelerate upwards at you in very dramatic fashion as you get close to landing.
Add to that that water is generally featureless. No bushes, fences, trees as you are coming down for you to orient on. Nothing but a flat surface with some waves -- and you don't have a frame of reference for how big those waves are. The result is that it is easy to cut away at 100 feet, or at 50 feet, or at 30 feet, just as you start to get into the "ground rush" zone and the water suddenly starts to look much closer than it did just a second ago. Dropping from 50 feet is a good way to go 20 feet under, and break or dislocate a hip or shoulder on the way, or to wind up hitting the water in some attitude other than feet-first, which can be lethal just in itself.
So yeah, bailing out over water was taught, and was important, but it was not a desirable way to end a flight.
Now as to ditchable vs. non-ditchable planes. The one plane that the USN operated (to my knowledge) that was non-ditchable was the PB4Y (navalized B-24). The USAAF, RAF, RN, and several other commonwealth air services also flew them for coastal / anti-submarine patrols. And the USAAF chose the B-24 as it's primary strategic bomber in the Pacific until the B-29 became available, so even when it was not doing over-water patrols it was often flying over water.
It was chosen because of it's long range, long loiter capabilities, and high payload. All very good. But it could not be safely ditched. The B-17 ditched quite easily and safely. But not the B-24.
The problems were two -- the bomb bay doors and the high wing. The high wing meant that the fuselage that came into contact with the water without that large, flat "surfboard" of the bottom of the wings to keep it skimming along. And the bomb bay doors were thin corrugated links, rather than flat panels.
The result was that if you set a PB4Y / B-24 down on the water, the two large bomb bays would cave it, and the crush of the water hitting the inside of the back half of the fuselage would break the bomber in half, before the wings even came into contact with the water. A mess, all in all. And one that did not lead to high survival rates.
So despite all that I said about the hazards of parachuting into water, PB4Y / B-24 crews were taught to bail out and let the plane go along it's merry way.
It was not a very popular plane with it's crews.
Kinda long-winded. Sorry 'bout that. But there were, indeed, good reasons for ditching or not ditching.
-Mark
(aka: Mk 1)