1992 Chevy Cavalier Loping idle in closed-loop, loss of pow

I have been trying to diagnose a problem in my wife's 1992 2.2L (VIN 4) cavalier RS automatic. I am going to be as detailed as I can, so please bear with me! The problem is primarily at idle. At times, the car will idle very high on startup, and at other times will idle very low, almost dying. More commonly, it will start and seem fine. Once the car has warmed up, the idle will lope, idling normal for about 5 seconds, then dropping low and almost stalling for about a second before jumping back up to a normal idle and then repeating.
Sometimes, the car will be running fine on the road, and the engine will start stalling out/sputtering, then about half a minute later, returning to normal. There are no vacuum leaks, and changing the idle air control and MAP sensor had no effect. No trouble codes are stored consistently, although once there was a coolant sensor too high, and another time there was an EGR related code (I can't remember which), although like I said, these are not repeatably stored (once each, I believe). Sometimes, the car appears to be idling okay, but when you press the gas (in drive), the engine has no power at all. Holding the pedal to the floor eventually leads to acceleration, but it lags by several seconds.
I began to suspect either wiring or the ECM, so I removed the glove box to get at it and probed each of the connections. Each seemed to have the correct voltage on it, so I removed the ECM connectors and gave both the connectors and sockets a spray with contact cleaner.
Very recently, I purchased a Auto X-Ray brand scantool (EZ-Scan 3000), and this verified that the sensors were responding as they should, but it turned up a strange parameter. The computer things that the car's battery has between 5 and 10 volts on it (it fluctuates), although the battery reads 12V with the car off, and it jumps to 13.8V with the engine running (so I know that the alternator is working). Swapping the battery made no difference in the reading. I cannot see anywhere on the wiring diagram for the car where the ECM gets information on the battery charge, so I assume that it gets this data from its power line, and this reads what it should (12V and 13.8V). I strongly suspect that the computer has a loose connection. I drove the car while monitoring the sensors, and when I stomp on the gas (and the car won't accelerate), the O2 sensor voltage drops to about 50mV, indicating a very lean condition. Perhaps the computer is not triggering the injectors because it is browning out (I hear the fuel pump momentarily run when I turn the key to ON, so I know that it is good)?
At any rate, I strongly believe that the computer is the problem here, and I am looking for some confirmation of this or a suggestion of what else could be the problem. If I need to replace the computer, is there anything special that I need to do? My manual shows a little metal flap, under which is supposed to be the car-specific calibration, but the ECM on this car has no such flap. Does this mean that I just need to swap out the computer? Thanks in advance (especially if you are still reading!)