One way to do this is to construct a corpus with these authors' writing as samples, then annotate the samples with the intented constructs that you desire to study, for example, annotate all cases of conceptual metaphors that you can identify, after you have done the annotation job, use a concordancer to extract all cases of conceptual metaphors, and compare their distributions (frequencies) across authorial differences, statistics, graphs, etc. may be used.

I believe that a quantitative analysis would mostly yield frequencies which might sometimes not be enough. I suggest an accompanying qualitative analysis in which the natures and themes of image schemas and metaphors are identified. For instance, horizontal, vertical, and containment image schemas may be used by different authors. The adoption of one of these carries implications for realty. Concerning metaphor, an eye for the nature of the source domain is useful. Various source domains could be adopted from the human, animal, food, etc. knowledge domains, and their effect on the target domain is bound to be different.

Are they even measurable? I do not think so. We might guess which process may have happened in his mind, so that he uses an especial world in a sentence or using a sentence in an especial way. We may say what might be involved, but to count them, I believe, will not be possible, as we do not know which experiences he had, which mental spaces he already has in his mind and the processes that has been done in all his lifetime that leads to using words, phrases or sentences in a special way.

you can categorize the writing into different scales such as knowledge of grammar , sentence structure, complecty of words , artistiic ect. don't you?

Use corpus tools . Annotation at the top of the list