Wednesday, January 17, 2007

Liberating On Liberty (from the Library)

Whew! I got some good linkage from my last post on ethicists stealing books, lifting this blog into the rarefied territory of almost thousand visitors a day -- peanuts for the likes of kottke.org and ordinary for Cognitive Daily -- but virtually a populist revolution for me.

The previous post about missing ethics books focused entirely on books first published in 1960 and later. I promised a separate analysis of classic, older texts. Here it is.

I looked at 10 classic (pre-1900) texts outside of ethics and 12 classic ethics texts. Selection criteria were a bit complicated; but please let me know, readers, if you think there are classic philosophy books I should have included on the lists below but did not include that are: (a.) influential, (b.) widely checked out, (c.) either clearly in ethics or clearly outside of ethics (not a blend), and (d.) generally published under separate cover (not as part of a general anthology including both ethics and nonethics texts by that author). The libraries I looked at were: UC Berkeley, UC Irvine, UCLA, UC Riverside, UC San Diego, UC Santa Cruz, Harvard, Stanford, Cornell, Columbia, Princeton, Michigan, and Texas.

I looked at missing books both as a percentage of books off-shelf (where "off shelf" means either checked out or missing/lost) and as a percentage of holdings. Data were collected over winter break so as to minimize effects due to texts being assigned for a particular class. (I will collect comparison mid-term check-out data in a month or so.)

The results, then, by book:
(Author, Title, Missing as % of off shelf, Missing as % of holdings)
Ethics:
Aquinas....Summa Theologica.......8.1%....1.2%
Aristotle..Nichomachean Ethics...11.6%....1.9%
Bentham....Principles of Morals..14.3%....1.4%
Hobbes.....Leviathan.............22.4%....3.6%
Kant.......Groundwork............22.0%....2.2%
Kant.......Metaphysics of Morals.12.5%....1.7%
Kant.......Second Critique.......12.9%....1.7%
Locke......Treatise of Gov't.....19.3%....2.3%
Mill.......On Liberty............30.4%....5.0%
Mill.......Utilitarianism........27.6%....2.2%
Plato......Republic..............19.7%....2.9%
Rousseau...Social Contract.......30.0%....3.4%
Total Ethics.....................19.4%....2.6%

Non-Ethics:
Bacon......New Organon.............3.3%...0.6%
Berkeley...Principles.............11.1%...1.0%
Brentano...Psychology Empirical....0.0%...0.0%
Descartes..Meditations.............6.1%...0.7%
Frege......Grundlagen.............29.4%...6.9%
Frege......Geach & Black trans.....0.0%...0.0%
James......Principles of Psych.....6.3%...1.7%
Kant.......First Critique.........10.1%...1.8%
Kant.......Third Critique..........3.6%...0.9%
Locke......Essay Concerning Human.16.3%...1.6%
Total Non-Ethics..................8.7%....1.4%

In sum, the classic ethics texts were about twice as likely to be missing, overall, than the non-ethics texts (p < .001; one-proportion tests, two-tailed).

A few books were outliers, with substantially higher checkout numbers (Plato [132] and Aristotle [95]) or substantially lower (Berkeley [18], Brentano [6], Frege [17 and 7], Bentham [21], and Kant Metaphysics of Morals [16]). (Instability due to small sample size may explain the highly variable results on some of the latter books, esp. Frege.) Excluding these books left short lists of books with very comparable holdings and checkout rates, and large enough holdings and checkouts for some statistical stability. Ethics books were still about twice as likely to be missing. (The odds ratios actually went up a bit.)

Looking by book, excluding those books with fewer than 25 off shelf, almost all the books with above-average missing rates are ethics books; almost all those with below-average rates are non-ethics. This effect was so pronounced that even with such a small sample of books (6 non-ethics, 10 ethics), the difference in the mean percentage missing was statistically detectable (7.6% vs. 20.4%, p = .001; 1.2% vs. 2.6% p = .005).

P.S.: I should add that due to the different types of courses classic ethics and non-ethics books are typically assigned for, and due to the high -- and perhaps not comparable -- rates of checkout by undergraduates, I think these data are not as revealing as the data described in my earlier post, which focused on a larger set of more comparable books, likely to be checked out only by professors and advanced students.

2 comments:

Brad C said...

Hi Eric,

This really is a good and fun topic. Someone else likely already asked this, but...

What does "off-shelf" mean?

I have a hunch that the character of the patrons checking the books out would most reliably track the relative value of (missing/(frequency of checkout)).

If that is not what you are calculating it might provide a defense of ethicists: perhaps their books are missing because they are more frequently checked out.

In any case, Wish I was in Sunny California!

Eric Schwitzgebel said...

Thanks, Brad, for your kind remarks, and for pointing out that I was unclear about "off shelf". I've added a parenthetical remark to the original post: Off shelf means checked out (including overdue) or missing.

I agree with your remarks, and the figures I'm looking at are missing/(checked out + missing); though I'm also looking at missing/holdings. Either way, it's not so good for the ethics books.

[By the way, I didn't calculate missing/(checked out not including missing) because in my earlier sample of books, this would have meant in some cases dividing by zero (when the only books off shelf were missing). Apart from that difference, the statistics turn out essentially the same either way.]