By Jonathan Rothwell - an author of the last one (at least from the summary - I haven't read the whole paper year) - here.
The word choice by Rothwell is a little confusing. For better or worse "STEM worker" means college-or-above science and technology worker. That's simply what people mean when they use the term. It often excludes things I think it would be nice if it did include, like health professionals, but you've got to use it in a clear way. With Rothwell's work here he's doing things like taking jobs that say they require basic algebra (see below for how he figures that out and the potential problems) and calling that a STEM job. That's very different from what most people think of as "STEM"... that just means you need to have done OK in maybe your freshman or sophomore year of high school because you might use it a few times on the job).
So Rothwell hitches on to the STEM bandwagon in a way that I don't think makes a whole lot of sense but in the process he makes a good point that I've been saying on the blog as well: there are a lot of good, technical, mid-skill level jobs in the U.S. and the bifurcated nature of our education system means that there's a lot of opportunities in these sorts of jobs for low-skill workers. What's required for them to take advantage of those opportunities is finishing high school, perhaps getting some career and technical education (CTE), apprenticeships, and community college. The other thing to note about these jobs is that they also offer the opportunity for further education and advancement. It's not uncommon at all for apprentices, after working in a mid-skill technical job, to then decide to go on to college and get higher level training in the field.
I am also a little leary of the O*NET data he uses to classify jobs... take that with a grain of salt. I've heard that a lot of the O*NET job characteristic stuff is dubious. But as another way of doing a big-picture look at the labor market I'm sure it's fine.