This is a classic case of Amdahl's Law[0], with square footage as the commodity being optimized rather than computational resources. The more "efficient" you make the scalable quantity (square footage directly allocated to each occupant in this case), The more dependent The final number becomes on the non-scalable terms (square footage of shared spaces like restrooms, stairways, hallways, eating spaces, and etc). Even a simple double loaded corredor with shared restrooms like a high density college dorm layout has a significant fraction of its base determined by the nonscalable spaces.

Nowadays anyone involved in high scalability computer processes is familiar with these concepts but in the 1960s when Amdahl first started focusing on the nonscalable processes the idea seemed as crazy as these square footage numbers seem to non-architects.

[0]https://en.m.wikipedia.org/wiki/Amdahl%27s_law

Another example of this is that many open office buildings that are (or seem) relatively high density are over 150 sqft/head because of restrooms, meeting rooms, and corridors.

Kitchens, administrative space, etc would be my guess.

> 235 square feet per person? That's a lot of room.

What? No it isn't.

For an application where you would expect shared spaces for kitchen/bath/etc, and a trade-off of comfort vs capacity, it is. It is higher than other family shelters. Which could be unused space, or perhaps colocated adminstration, or whatever. It wasn't a complaint, just an observation.

For a comparison, at my university dorms its roughly 65 square feet per person.

Do you mean that your room is roughly 10'x13', shared with one roommate? If so, you're not comparing apples to apples. Amazon's number is derived from the entire building (including common hallways, stairs, bathrooms, and other spaces). Also, while a dorm room allocating 65 sqft per person might fit a loft bed, desk, and a student's possessions, that's not convenient for a family shelter.

235 sq ft per person seems fair. Ikea has a 400 sq ft setup for 3-4 people.