The fact that many of the best selling books of all time were published a long time ago should be no surprise even without the supposition that people used to be more literate. The number of books published back then was much, much smaller. A book's distribution is likely to be better when it's one book out of seven to come out that year, rather than one book out of three thousand. Note that the link you provided doesn't even attempt to argue that literacy rates were as high back then as they are now, it acknowledges the opposite.

I'm not complaining that you didn't make a full argument, I'm complaining that the Harpending link was worthless. It would be better if you had omitted it. Provoking people with bullshit does not suggest that you are interested in the truth. At least Harpending has the humility to admit when he is wrong. Like you, he assumed that if he drew an example at random it would support his case, but he had the humility when presenting it to ask for analysis of the particular case. The fact that he and you are wrong in the belief that randomly pulling examples supports you is evidence that you are overconfident. [...] Well, if you are familiar with the literature, why not cite a survey or something? After the Harpending example, why should I trust your description of the literature? But thanks for the links in this comment. Actually, I like simple arguments like Cochran's much more than the education research literature. Mainly I distrust it because it is politicized and incompetent, but on this particular point, I imagine that most of the (US) literature is worthless because (US) schools lie about their spending. It is widely known that poor urban districts have larger budgets per student than rich suburban districts, but this is an illusion because suburban buildings are off budget, especially in the richest districts. Probably some of the literature, like the nameless commenter, isolates instructional spending, which is a much more reasonable comparison, in addition to evading the deceit.