Technically, the best protection is a self-contained breathing apparatus (SCBA) which is a fancy way of referring to a respirator connected to an oxygen tank, but that thing too impractical and overkill for most people.

The amount of protection offered by a positive air pressure respirator (PAPR) depends on what kind of hood is being used and may offer about the same (or more) protection than a reusable elastomeric respirator. Assigned Protection Factor (APF) is a measure of the level of protection offered by types of respirators; PAPRs range from 25 to 1,000 APF, whereas reusable elastomeric respirators offer an APF of 10 (probably around 25 in reality) for half-facepiece respirators and 50 for full-facepiece respirators. Besides a potentially higher APF of 1,000, PAPRs also don't require a fit test (which most people aren't going to bother doing), so leaks are less likely, even though the APF might be similar or even less than a full-facepiece respirator. They're also more comfortable than any other respirator. However, one hard-to-avoid disadvantage of a PAPR is that you'll have to carry around an elastomeric respirator as a backup.

To mitigate the cost and bulk issues, it's possible to DIY a PAPR (a plastic bag connected to filters, fans, and a battery). A DIY PAPR might be easier to repair and could also filter exhaust air, unlike commercial PAPRs. I haven't bothered to DIY it yet, but others have definitely used these things successfully.

DIY PAPR intro info

https://viralhelmets.medium.com/15-viral-helmet-version-of-a-2000-hospital-mask-papr-4950905ae2cc
https://www.youtube.com/watch?v=zj_C4GrxfNM&t=102s
https://www.youtube.com/watch?v=xfaswOLIHoQ

DIY PAPR in-depth info

https://www.designnews.com/industry/beyond-n95-hackathon-produces-air-purifying-respirator
https://devpost.com/software/bunnypapr-for-jcrmrg-hackathon
https://bunnyscience.dozuki.com/Guide/Bunny+Science%E2%84%A2+PAPR/4
https://web.archive.org/web/20210202224805/https://bunnypapr.org
https://web.archive.org/web/20210825175428/https://www.viralhelmets.com
https://www.instructables.com/Viral-Helmets-9999-Hospital-Grade-Viral-Protection/
https://www.reddit.com/r/viralHelmets/

I think it would be a lot cheaper and less conspicuous to put a really cheap scuba kit under a cloth mask/bandana and a normal backpack, so that only the tube going between them is visible. Easily 1/3 the price too. Maybe including swim goggles or airtight lab goggles, I haven't found any info on how likely the virus is to enter the eyes via respiratory droplets.

One way or another, the best thing to do is to be careful in apartment common areas, and to meet people outside, especially if eating takes place.

SAT decreases due to school disruption seem pretty likely

I'm really sorry that I can't go into details here, but that's exactly the kind of information that you'll have a hard time finding in the modern era (without paying a handsome sum of money to a well-connected, exclusive firm you've actively established a relationship of trust with).

If that kind of problem ends up routinely causing you significant trouble in the future, we can eventually talk about it via a call. SAT trends in particular are harder than they look, especially if they start using/already have been using different question combinations/ordering for different people, which I predict will happen by 2027 if it hasn't already. I also predict that publicly available SAT overviews will either become less available, less representative, or otherwise significantly less accurate, in the same timeframe.

It would be helpful to hear more specific thoughts on whether / why this data is "garbage".

An initial thought on that (since I don't have time to dive deeper on this today) is that the first source linked in this post says most of the studies did not use a control group:

For the assessment of the quality of the included studies, we used the Newcastle-Ottawa Scale for observational studies.(22) Not every item from this scale was relevant for the included studies as most included studies did not use a control group. Therefore, we used two items from Selection (representativeness and ascertainment of outcome) and the three items from Outcome (assessment of outcome, follow-up long enough, adequacy of follow-up; see Supplementary Material).