One thing that has not been mentioned is that an increasing number, especially larger-size sensors is what is called "ISO invariant" these days, meaning that with regard to the raw sensor data, raising ISO and lowering +EV are the same: either will negotiate for less exposure. Raising ISO will, however, in the subsequent processing then amplify digitally the lower exposure (with smaller and/or older sensors, part of the amplification is in the analog domain).
That means that ISO on these cameras does nothing for the raw data in M mode: it just changes the interpretation of it when converting to JPEG. On other modes it influences exposure. If you have a strongly backlit image and use +EV for bringing the subject up (like birds before the sky) in the JPEG, you might want to avoid overexposure, namely blown highlights. You can do this by raising the ISO above base ISO. If you use +2EV while metering on a larger area (spot metering is just too fragile) and your camera has about 1EV of reserves before blowing highlights in raw, you might want to double ISO from the base ISO in order to reduce exposure while not changing the brightness from +2EV.
Of course, with birds before a bright sky, a polariser may also help lessening the contrast between bird and sky, assuming that the light is sufficient.
At any rate: for exposure corrections of amounts like +2EV, consider raising the ISO as well to avoid blown highlights. There may be a problem with some camera's JPEG processing, however, that the camera takes the higher ISO value (regardless of the amount of +EV you set) as license to overindulge in noise reduction, smearing colors in the process. So retaining raw files will make this safer, and setting JPEG noise reduction in-camera to conservative or minimal settings might also be a good idea.