Stop talking about "bias" and "assumptions". I've made no such claims regarding your work; I'm addressing it point-by-point, and I expect you to do the same with me, without the accusatory language. I'm analyzing data here, with no preconceptions about whether Bob is in the suit or whether it's someone else, or indeed whether the P-G subject is a large hairy person, with no suit or costume evident at all. Let's stick to the facts, and identify opinions as such, shall we? Thanks.
Subjectively, the top-line of the P-G subject's head could come down a fraction, to where the horizontal slope of the "brow" reaches a plateau before coming up again. This would narrow the distance between the top-line and the eye-line on the P-G subject.
The oblique tilt of Bob's head, whether it's precisely as I've indicated with the admittedly crude MS Paint or whether it's closer to your own recent, slightly less oblique indication, means that there is a greater distance between the top-line of Bob's head and his eye-line than there would be if his head were tilted more vertically. IOW, this distance will narrow if you pull his head straighter. If we could somehow adjust the angle of the head to be more in-line with the P-G subject's, we could then re-size the images to match up more evenly.
In short, the two pics do not lend themselves to a 1:1 comparison.
You can form all kinds of opinions about them as they are, but owing to the numerous discrepancies between head placement, camera lens size, angle of the camera-man, pictoral compression by the processing software, and limb position, it's practically impossible and virtually worthless to compare them or to offer opinions about those comparisons.
Such an exercise in visual data analysis is fun and challenging in its own right, but gets us nowhere with regard to determining the identity of the P-G subject.