Ensuring high quality sources where needed, especially science/academic topics (narrow focus)

I think those are almost two different problems.

One class of article is low quality because it was substantially produced by one or two editors, with everyone else just adding a random statement at a time. No one has ever thought about researching it or organizing it or focusing it on what is relevant. The other class of article is low quality because there's too much activity -- it's constantly attracting new editors who all have an opinion, not to mention the cabals who have a stake in keeping out anything they disagree with. If we had four recommendations, I might spend two on trying to improve the stubs and low activity articles, with another two recommendations just for the high activity articles that suffer from controversies.

I might even focus a little more on the stubs and low activity articles, and put three recommendations on those. First, it's more widely desired in the community. People might disagree about how to resolve the Israel/Palestine or Race/IQ articles, but most good editors will agree that an unsourced article is a bad thing. Secondly, a controversial article will at least be well written. The dispute may let systemic bias slip in, but you'll find that these articles are pretty well written with a lot of sources. The low-activity articles are unmonitored, meaning you either get stubs that haven't had substantial work in months, or articles where people add all kinds of garbage with no effort to monitor it for quality control.

The low activity articles need peer reviewed sources. That would be a huge step up from the unsourced stuff, and the stuff sourced to blogs and personal websites.

The high activity articles need something else. When you have a well written article that has sources, but it still suffers from major problems with bias, you need an authority to step in. An expert. A mediator. An arbitrator. An editorial board. A neutral third opinion.