No, the major goal is to examine the validity, integrity and contribution of the work. Since these can be challenged, causing the work to be rejected or subjected to a major revision, editing at this stage would just be like staining the wood before the carpentry work.
Also, at least for me, my mind runs on two different gears when engaged in editing and reviewing. If I have to edit, I can only edit; if I have to think about the concept, I can't edit... perhaps in my mind editing only happens when the concept is already there.
Having said that, there are three areas I always do a full body pat down, once for concept, once for edit:
- Table: I suggest always comb through the table title, column and row titles, numeric agreement, footnotes, superscripts, etc.
- Illustration: Check labels, title, and footnotes.
- Abstract: Look for typos and point out if a particular sentence appears to be terse or even misleading.
The reason for putting the abstract under all lenses is apparent. The reasons for checking the illustration and table are that these are usually the most neglected places in copy editing and revision. And yet, these are the most viewed sections in a paper. Nine out of 10 errors I've spotted in journal articles are in these two hot spots.
In the main text, if there is any sentence that is very grammatically challenged, I cite the page and line in the review, and comment on the need of a rewritten version.
If there are typos, I collectively write one separate comment requesting a through spell-checking, and give 2-5 occasions in the article as examples.