The logic behind it is that if you’ve done a sloppy job with the presentation of your arguments, it becomes likelier that you’ve also cut corners during the research phase. It’s not about the quality of the prose, but about the lack of effort for something so easily fixable (a software spellchecker will catch the most egregious mistakes).

I’m not aware of any hard data on the matter, but as you say, it seems like most people are concerned by this and yet the author did nothing about it, further cementing the careless attitude.

To be perfectly clear, my comment is a generalisation to answer your point, not a critique of this particular author or their work (I haven’t read it).