They had to pick and choose what to go into depth on, but when they wanted to go into depth, they would hire the best writers in the field to explain everything really well. Issac Asimov penned more than one encyclopedia article, as did other science fiction authors.
I honestly learned most of my fundamental science knowledge from old 1950s and 1960s science fact books. Asimov in particular has multiple books that explain science really well, down to the level of nuclear fusion and fission, biochemistry, and a lot of other subjects, and I had a few other books in the same vein that were all really old, and really good.
I remember one book that was biographies of famous scientists, it went into depth about Marie Curry and her contributions to science. Reading that ~age 10 certainly impacted my views on gender equality and helped me believe that potential is not limited by gender.
Wikipedia, for all the great things about it, won't really do that. There isn't a Wikipedia page I can point literally anyone[2] at and say "read this and you'll understand the fundamentals of radiation and atomic physics".
I am sure such resources are out there on the web, but the odds of stumbling upon those resources at random is less than the odds of coming across a good science book at random in a thrift store.
That said there are some awesome YT science channels who have worked to fill this gap in quite well, but video sources are different than written sources, with each having their strength. (I read that short 10 or so page biography of Marie Curry probably a half dozen times, not going to do that with a YT video!)
[2] Where anyone is defined as "age 10 and up, with no mathematical background, who wants to be entertained while they read."
That's the entire point. I'm not the expert. Nor am I able to even understand the basics of which direction to read/learn - that's what the expert curators are for.
I dont have source at moment. But I remember reading a few years ago that wikipedia had fewer errors than most all encyclopedias on average.
There is definitely something to be said about consistency and clustering but on average I believe wikipedia is still the best in fewest errors.