It would be helpful to have other measures such as the median and the spread in scores. In particular, if the high end of the natural spread gets cut off, the a

It would be helpful to have other measures such as the median and the spread in scores. In particular, if the high end of the natural spread gets cut off, the average can be skewed as there is more room below 6 than above it in a 10 point scale. The median, quartiles, and standard deviation would help to alleviate those concerns. As a quick example, compare a spread of 2: 6,4,8 average 6 versus a spread of 6: 6,0,10 average 5.3ish where that 10 ought to be a 12 given some underlying real variability, but it gets cutoff by the upper limit of 10. As for the claim of neatness, I can attest at having graded math for over 20 years that neatness of work is very impactful, particularly for partial credit. I have often caught myself just before giving different scores for essentially the same argument based on presentation. The variability in presentation can be immense, even at the post-college level. And the work often, though certainly not always, conforms to sex stereotypes. This applies most to handwritten work, but even typed up work shows this variability in its organization and depth.