If you're in the Bay Area, I invite you to drive around on Middlefield Rd in the 25mph sections to see how well setting speed limits way below what most drivers would consider reasonable works in practice (most people travel ~35-45mph there).

Further reading if you're really curious about this situation in particular, there's been a lot of back and forth around Menlo Park / Atherton specifically: - https://www.ci.atherton.ca.us/DocumentCenter/View/8751/Engin... - https://www.almanacnews.com/news/2020/10/22/menlo-park-counc... - https://www.paloaltoonline.com/news/2016/11/17/palo-alto-eye... - https://www.cityofpaloalto.org/files/assets/public/agendas-m... - on 85th percentile theory: https://dot.ca.gov/programs/safety-programs/setting-speed-li...

Particularly amusing is https://www.mercurynews.com/2016/11/30/radar-isnt-best-optio... ... where most people seem to agree that everyone drives too fast, but in order to use radar / lidar to measure speed, they'd need to comply with state law which would mean /raising/ the speed limits in order to comply with the 85th percentile rules, so instead they just stick their heads in the sand and continue to believe, despite many studies finding to the contrary, that zero enforcement and a sign that says 25mph will actually get people to drive on 5-lane roads at ridiculously slow speeds. This leads to an exciting mix of a few cars trying to actually drive the 25mph speed limit while most other people try to frantically pass them. Extra fun on sections that don't have left turn lanes, where the left of the two lanes becomes the de-facto turn lane, then combine that with the general lack of turn signal usage and you get...a very frustrating commute indeed. But what do you want for your $3m 1000sqft single-family home in those neighborhoods??

Edit: there's a lot more traffic engineering theory around speed limits vs how fast people actually drive. The last couple of points in particular are extra fun: https://www.saratoga.ca.us/DocumentCenter/View/998/Setting-S...

    The rationale for speed zoning is also based on the following fundamental concepts, which are
    deeply rooted in our system of government and law:
    •  The normally careful and competent actions of a responsible person should be considered legal.
    •  Laws are established for the protection of the public and the regulation of unreasonable
    behavior by the individual.
    •  Laws cannot be effectively enforced unless the majority of people voluntarily consent to and   
    comply with them.
 
    In general, most people accept, understand, and follow these concepts. But, as mentioned
    earlier, when confronted by local traffic challenges—an increase in traffic volumes or the frequency
    or severity of car crashes in a particular area, for example—people often reject these concepts and
    rely instead on a number of widely held misconceptions, such as:
    •  Lowering a posted speed limit will reduce traffic speeds.
    •  Raising a posted speed limit will increase traffic speeds.
    •  Lower speed limits will reduce the rate of crashes and increase safety.
    •  Any posted speed limit is safer than an unposted speed limit.
 
    In fact, these commonsense notions are not supported by studies, which consistently
    demonstrate that:
    •  The speed of traffic does not change significantly after new or revised speed limits are posted.
    •  There is no direct relationship between posted speed limits and the frequency of traffic
    crashes.

Lane width vs speed is interesting to read about too: https://nacto.org/docs/usdg/review_lane_width_and_speed_pars... (tl;dr people will naturally drive more slowly on streets with street parking, narrow lanes, curves, driveways, visible foot traffic, etc, and drive more quickly when roads are visibly more open)