In order to be comparable, not the field of views should be the same but the size of the objects in pixels. Then the larger image should be cropped to the size of the smaller one, and then it becomes reasonable to compare the effect.
Surely it depends on what you are trying to compare, Gabor. Most photographers are concerned with the appearance of the final print (or image on display monitor); is it sharp and vibrant; does the file have enough resolution to blow up large; is there good detail in the shadows or, is there objectionable noise in the shadows; is the tonality smooth and natural etc?
If these are your concerns, then it's essential to compare images with the same FoV since the FoV is central to the composition. Also, if the sensor is a different size then it's essential to adjust f stop to achieve equal DoF in both images because degree of DoF is also central to the composition. Choice of f stop also affects resolution.
However, there are good reasons to compare 'pixel for pixel' performance and that certainly can be useful information to have. For example, if both my longest lenses for MFDB and 35mm were 300mm, I might want to know whether I should use a 1Ds3 or a P45+ for a particular shot, if the composition required a 35mm size and aspect ratio.
In other words, will the P45+ image with 300mm lens, after cropping to the same FoV as the 1Ds3 shot with 300mm lens, be better quality? In order to predict which is likely to be better, I need to compare 'pixel for pixel' quality, as well as lens quality.
I would suggest that most photographers try to maximise the benefits of their sensor 'real estate' and choose a lens and perspective which result in the least amount of cropping.