This is an old thread but it's the most prominent one on this test.
EQ 3
SQ 47
I have looked at an some explanations of how the scoring works and the test gives a fair amount of scope for anomalies. It's far more accurate than guessing at random but still probably is within an appreciable margin of error. They have clearly done some additional things as well to scoring that I have not been able to find a written explanation of. I assume however that certain questions might have different weights and that some might exclude others.
There are problems where a non-strong answer is ambiguous. It appears the stronger the agreement or disagreement the stronger the score. This can be problematic where people are either more cautious in answering or where questions are not mutually exclusive. By that I mean that for some questions, being closer to the middle may either be because you're bad at both things or good at both things, it is ambiguous.
There are a few versions of the test out there and some have (strongly agree, agree, disagree, strongly disagree) where as others have (strongly agree, slightly agree, slightly disagree, strongly disagree).
You can play around with the scores to attempt to compensate for some of that.
A really simply way is to turn your scores into ratios, here are some examples:
Total 50
EQ 3 -> 6
SQ 47 -> 94
Total 77
EQ 44 -> 57
SQ 33 -> 43
Total 47
EQ 25 -> 52
SQ 22 -> 48
Total 46
EQ 42 -> 91
SQ 4 -> 8
Total 47
EQ 31 -> 66
SQ 16 -> 34
Total 130
EQ 94 -> 72
SQ 36 -> 28
On a random side note I have discovered that this research proves that guys with Asperger's Syndrome are real men.
That was intended humorously.