RUNNING HEAD: Do Ethicists Steal More Books?
Do Ethicists Steal More Books?
Eric Schwitzgebel
Department of Philosophy
Do Ethicists Steal More Books?
Abstract
If explicit reasoning about morality promotes moral behavior, as Kohlberg and many ethicists have suggested, then one might expect ethics professors to behave particularly well. However, professional ethicists’ behavior has never been empirically studied. The present research examined the rates at which ethics books are missing from leading academic libraries, compared to other philosophy books. Study 1 found that contemporary (post-1959) ethics books were actually about 25% more likely to be missing than non-ethics books. When the list was reduced to the relatively obscure books most likely to be borrowed exclusively by professional ethicists and advanced students of ethics, ethics books were almost 50% more likely to be missing. Study 2 found that classic (pre-1900) ethics books were more than twice as likely to be missing as other classic philosophy books.
Keywords: Morality, ethics,
philosophy, reason, moral reasoning, Kohlberg
Do Ethicists Steal More Books?
In the Nicomachean Ethics, Aristotle writes that the aim of studying ethics, “is not, as in other inquiries, the attainment of theoretical knowledge: we are not conducting this inquiry in order to know what virtue is, but in order to become good” (4th c. BCE/1962, 1103b). In Groundwork of the Metaphysics of Morals (1785/1998), Kant asserts that a philosophical understanding of morality is essential to a firm and unswerving hold upon morality, allowing one to avoid corruption by self-regarding desires and misdirection by irrelevant considerations. In On Liberty (1859/2003), Mill says that ethical doctrines become largely powerless to guide behavior unless their fundamental principles are grasped and the arguments both for and against them are rigorously considered. If Aristotle, Kant, and Mill are right, a lifetime’s commitment to the study of ethics should have salutary effects on one’s moral behavior. Philosophical moral reflection is not, or should not be if done properly, morally inert. Ethicists, by virtue of their long study and deep appreciation of the principles of morality, will behave better than non-ethicists. Even if the study of ethics is not itself morally beneficial, one might reasonably expect that people drawn to a career teaching and studying ethics will, on average, be more deeply committed to the importance of morality than those drawn to other careers either within or outside of philosophy.
But do ethicists actually behave better than non-ethicists in philosophy, or than non-philosophers of similar social background? The question has never been systematically studied.
Among contemporary philosophers, Nussbaum (1997) and Moody-Adams (1997) have argued that philosophical moral reflection is a crucial part of our moral education, will tend to make us better citizens, and can give us the courage to stand up for justice. Though their arguments have a kind of a priori plausibility, they offer no empirical evidence in support of these claims. Posner (1999), a judge and philosopher of law, argues in contrast that academic philosophy as normally practiced “has no prospect of improving human behavior” (p. 7) since (among other things) its motivational force is feeble and it tends to traffic in rationalizations either of self-interest or of culturally derived, arational intuitions. Opinion among ordinary professional philosophers appears to be divided. Joshua Rust and I (Schwitzgebel and Rust, in preparation) polled passersby at an American Philosophical Association meeting in April 2007 and found that although a substantial minority (especially among ethicists) expressed the view that ethicists do behave morally better on average than non-ethicists, a majority of respondents said that ethicists do not behave better. In fact, a substantial minority of non-ethicists suggested that ethicists behave on average worse than non-ethicists. Anticipating the current research, several philosophers have shared with me their impression that ethics books are missing from libraries more often than other philosophy books.
In psychology, accounts of moral development that give a central role to moral reasoning, such as Piaget’s (1932/1965) and especially Kohlberg’s (1984), should, it seems, predict that ethicists will show excellent moral behavior, since there is no doubt that professional ethicists are champions of moral reasoning – at least in the sense of “moral reasoning” researchers like Piaget and Kohlberg have in mind. (Accordingly, Rest [1993] found that graduate students in moral philosophy show very sophisticated reasoning about Kohlbergian moral dilemmas.) However, the nature and strength of the relationship between Kohlbergian measures of moral reasoning and real-world moral behavior remain unclear (Blasi, 1980; Kohlberg, 1984; Colby & Damon, 1992; Krebs & Denton, 2005; Stams, Brugman, Deković, van Rosmalen, van der Laan, & Gibbs, 2006). In contrast, accounts of moral development and behavior that emphasize emotion and downplay the importance of reasoning (Haidt, 2001; Greene & Haidt, 2002), might predict that skill in philosophical reasoning about ethics would have little bearing on actual moral behavior.
Although the relationship between moral behavior and a professional career in ethics remains unstudied, a number of researchers have examined the relationship between professed religiosity or religious behavior (such as church-going) and self-reported or objectively measured social deviance or criminality. An influential study in 1969 (Hirschi & Stark, 1969) found no relationship between high school students’ self-reported church attendance or belief in an afterlife and either police records or self-reported criminal behavior. Subsequent studies have tended to find weakly negative (often non-significant) relationships between religiosity and criminality or moral deviance. A recent meta-analysis (Baier & Wright, 2001) puts the median effect size at r = -.11, but even this small correlation is likely to be inflated by the file-drawer effect (the fact that negative results, especially of small studies, are less likely to be published) and bias on the part of religiously committed subjects (who may suspect the hypothesis and unwittingly adjust their answers) and investigators. Also, many of the studies included in Baier & Wright’s meta-analysis are only correlational and control for only a few potential confounding variables, and thus give insufficient basis for inferring a negative causal relationship between religiosity and immorality or crime. Two recent studies that attempted to get at the issue of direction of causation through temporal analysis found results in the opposite of the predicted directions (Eshuys & Smallbone, 2006; Heaton, 2006).
The most closely related issue on which there are considerable empirical data is the effectiveness of courses in business ethics. Research in this area generally examines the relationship between business ethics courses and either self-reported attitudes about business ethics or maturity of response to Kohlberg-like dilemmas. The results are mixed, with some studies finding that students exposed to business ethics show more ethical or more mature responses (Boyd, 1981; Glenn, 1992; Hiltebeitel & Jones, 1992; Murphy & Boatright, 1994; Loe & Weeks, 2000; Luthar & Karri, 2005) and others finding a very limited relationship (Duizend & McCann, 1998; Conroy & Emerson, 2004) or none at all (Wynd & Mager, 1989; Borkowski & Ugras, 1992; Smith & Oakley, 1996; Martin, 2007). Many of these studies, unfortunately, lack control groups or control questions. Without control questions, students can be rated as “more ethical” by means of simple strategies. For example, a number of studies simply measure the degree of students’ self-reported condemnatory attitudes about hypothetical violations of ethical standards (either before and after a business ethics course or in comparison with another group). Students may then appear more ethical simply by showing a bias toward regarding any presented scenario or behavior as ethically problematic – a response strategy that ethical training courses may tend to encourage but which need not show any real improvement in moral understanding. I have been unable to find any empirical studies attempting to relate business ethics instruction and real-world moral behavior.
The Present Research
The present research aims to provide the first small bit of non-anecdotal, empirical data directly on the question of whether professional ethicists behave any better than non-ethicists. In particular, the present research examines the theft and negligent treatment of library books. Are ethics books more, or less, likely to be missing from academic libraries than other philosophy books? I choose other philosophy books as the comparison class, rather than books in literature, chemistry, or some other discipline, and rather than the whole universe of non-ethics books, because it seems likely that by looking only within philosophy, the patterns of book use and the population using the books will be more closely matched, reducing the potential role of a variety of confounds.
I assume that the theft and negligent treatment of library books is, as a general rule, morally bad. Whether a book is missing for innocent reasons or instead due to negligence or theft cannot be determined from library records, but no such assessment is necessary if we assume that the rates of completely innocent loss are the same between the two groups. (I will return to some potential problems with this assumption in the conclusion.) Furthermore, most loss is not entirely innocent, I suspect: One can handle library property with more or less respect and care, conscientiously keep library books separate from one’s personal property or mix them in. Some professors and graduate students, when they change universities, negligently or intentionally bring some library books with them. Even if a book is missing simply because it was misshelved by a patron without having been removed from the library, that reveals negligence and a disregard of most libraries’ requests that patrons not reshelve books. Books specifically described in library records as having been paid for by the patron were excluded from analysis. Not all libraries display such data in their online records, but among those that do, books recorded as “lost and paid” constituted a small proportion of the missing books.
Accidentally misshelving a book, choosing not to return a library book one notices after a move, not bothering to check out a book one has accidentally walked out with – these and the like are of course only minor faults in the grand scheme of things, and I myself can’t claim complete innocence. Even if (per impossible) the results of the present research were completely decisive and free of confounds, it would be absurd to draw general conclusions about the overall moral character of ethicists from such limited data (especially in light of situationist findings in social psychology: Ross & Nisbett, 1991; Doris, 2002). The research reported here is intended as no more than the first of a number of projects that will explore the moral behavior of ethicists in a variety of different ways.
Two studies will be presented. The first examines the rates at which contemporary, mostly technical philosophy books are missing from academic libraries. The second examines the rates at which classic pre-20th century texts are missing.
Study 1: Rates at Which Recent Technical
Philosophy Books Are Missing
Method
Compiling the list of titles. To generate the list of titles to be examined, I drew from two sources. First I compiled as list of book reviews published in Philosophical Review (a leading philosophy journal) from 1990-2001. Based on book title and author, an independent coder – an ethics professor – coded the reviewed titles as either (1) ethics, (2) non-ethics, or (3) marginal / don’t know. The coder was instructed to treat philosophy of law and political philosophy as ethics and philosophy of religion and philosophy of action as marginal. All titles rated by the coder as ethics (147) were added to the ethics list and one-third of the non-ethics titles (155) were added to the comparison list; marginal titles were excluded. Most titles were books published in 1988-2000, but some were published earlier and re-issued in a new edition or were new translations of older texts. None of the reissued titles were originally published before 1960, and none of the new translations were of books widely known outside of professional philosophy (the focus of Study 2). To supplement these lists I turned to a second source, the October 2006 edition of the Stanford Encyclopedia of Philosophy (SEP). I coded SEP entries as either (1) ethics or (2) philosophy of mind and/or language (the only subfield of philosophy comparable in size and scope to ethics). Titles were added to either the ethics list or to the comparison list if they were published after 1959 and appeared in the bibliographies of at least five encyclopedia entries in category (1) or (2). Seventy-six ethics titles and 63 non-ethics titles (many, but not all, in philosophy of mind or language) met the criteria, for a final list of 206 ethics titles and 213 non-ethics titles, with a tilt toward philosophy of mind and language among the non-ethics titles.
Most of the titles on both lists are virtually unknown outside of philosophy. The full list of titles is available at ****. Among the most widely known ethics titles are A Theory of Justice (Rawls), Concept of Law (Hart), and In a Different Voice (Gilligan). Among the most widely known non-ethics titles are Word and Object (Quine), Consciousness Explained (Dennett), and Naming and Necessity (Kripke). Relatively better-known titles – defined as those appearing on the SEP list, regardless of whether they also appear on the Phil Review list – were excluded from some analyses.
The lending libraries. I examined online status information for
every copy of these titles in 13
Collecting the data. A research assistant and I collected the ethics and non-ethics data in tandem between October 17 and December 13, 2006, when all schools were in fall session. We coded each copy of each title as either on shelf (neither checked out nor missing), within period (checked out but not overdue), overdue (overdue by no more than one year), or missing (including books more than one year overdue). In some cases, I called libraries to clarify the significance of their status messages. The following composite categories were defined for the purpose of analysis: delinquent (overdue or missing), off shelf (within period, overdue, or missing), and held (on shelf, within period, overdue, or missing; that is, all categories combined). I excluded from analysis books in repair, books listed as lost and paid, and books whose status was unclear.
I will use the term books (or sometimes, for additional clarity, copies of books) to refer to individual copies of books (e.g., “I excluded books in repair”) or collections considered as collections of individual copies (e.g., “8.7% of all ethics books”). When I intend to refer to a class of books sharing the same title and author or aggregates of such classes, I will use the term titles (e.g., “I examined 206 ethics titles” and “a title-by-title analysis”).
All statistical tests are two-tailed with an alpha level of .05 for significance and .10 for “marginal” significance.
Results
Raw numbers and key percentages are listed in Table 1. Analysis suggests that ethics books are more likely to be delinquent or missing than non-ethics books, both when considered as a percentage of holdings (two proportion test, 3.4% delinquent vs. 1.9% delinquent, 95% CI for difference 1.1%-1.9%, p < .001; 2.2% missing vs. 1.3% missing, 95% CI for difference 0.6%-1.3%, p < .001) and when considered as a percentage of books off shelf (two proportion test, 13.4% delinquent vs. 10.5% delinquent, 95% CI for difference 1.1%-4.7%, p = .002; 8.7% missing vs. 6.9% missing, 95% CI for difference 0.3%-3.2%, p = .02). Ethics books are also more likely to be simply overdue than non-ethics books (as a percentage of holdings 1.2% vs. 0.7%, two proportion test, 95% CI for difference 0.3%-0.8%, p < .001; as a percentage of off shelf 4.7% vs. 3.5%, two proportion test, 95% CI for difference 0.1%-2.3%, p = .04). Among these measures, missing as a percentage of books off shelf is probably the most valid, since it takes into account the popularity of the books and since missing rates are less confounded by mere absent-mindedness than delinquent or overdue rates. It is also a statistically conservative measure in this context, since its effect size is the smallest and its p-value higher than any but overdue as a percent of off shelf. Therefore, further analysis and discussion will center on this measure. The odds-ratio of missing as a percentage of off shelf (8.7% for ethics, 6.9% for non-ethics) is 1.25 to 1. In this sample, an ethics book is 25% more likely than a non-ethics book to be missing, if it is off shelf.
----------------------------------------------------------------------------------------------------
Insert Table 1 about here
----------------------------------------------------------------------------------------------------
Unfortunately, there are several potential problems with this analysis, pertaining to (1) differences in the average age of the books, (2) differences in the popularity of the books, (3) the fact that ethics books are used more in the study of law, and (4) the possible failure of independence between trials. The remainder of this section will re-analyze the data on missing as a percentage of off shelf, with these four concerns in mind.
Age of the books. As it happens, the ethics books are older, on average, than the non-ethics books. The mean copyright year of a within period (off shelf but not overdue or missing) copy of an ethics book is 1987.5, while the within period non-ethics books have a mean copyright year of 1989.7. And older books, as one might expect, are more likely to be missing, since they have had longer to disappear from the shelves. Merging ethics and non-ethics, the median copyright year of an off shelf copy of a book in this data set is 1990. 10.0% of off shelf books with a copyright before 1990 are missing, while only 6.4% of those copyrighted 1990 or later are missing (95% CI for difference 2.1%-5.0%, p < .001).
A linear regression model to correct for age proved to be unstable, with too much of the effect driven by a few very popular titles from the 1960s and 1970s, so instead the data were reduced, excluding from analysis all titles copyrighted before 1985. This resulted in a mean copyright year of 1992.9 for the within period ethics books and 1992.7 for the non-ethics books. On this reduced data set, the percentage of books off shelf that are missing is 7.7% for ethics and 5.7% for non-ethics, for an odds ratio of 1.35 to 1. The result remains statistically significant (95% CI for difference 0.4%-3.4%, p = .01).
Popularity of the books. A similar concern can be raised about the popularity of the books. As is evident from Table 1, ethics books are held and off shelf in greater numbers than the non-ethics comparison books. And the more popular books, as measured by number of within period (non-delinquent) checkouts, are more likely to be missing than the less popular books: Popular books are not only (by definition) more likely to be off shelf but also a higher percentage of those off shelf are missing. This can be seen by a median split: 49.2% of within period checkouts are of copies of titles with no more than 23 within period checkouts across all libraries. Pooling all copies (ethics and non-ethics) of these less popular titles, missing as a percentage of off shelf is 6.8%. Among the more popular titles, missing as a percentage of off shelf is 9.4% (95% CI for difference 1.1%-4.0%, p < .001).
I took recourse to further reducing the data. By excluding all titles listed in the Stanford Encyclopedia (regardless of whether they also appeared in Phil Review), the best-known and most checked-out books were all eliminated, leaving only relatively obscure books – which in any case are the books most likely to be checked out exclusively by professors and advanced students in philosophy and thus might be the best measure of professional ethicists’ behavior. After the exclusion, the book list contained 128 ethics titles and 149 non-ethics titles (alphabetically, the first four ethics titles were Against the Ethicists, Sextus Empiricus, trans. Bett; Alexander of Aphrodisias: Ethical Problems, trans. Dooley; Altruism, ed. Paul; and Am I My Parents’ Keeper?, Daniels; the first four non-ethics titles were Abstract Objects, Hale; Alfarabi, Avicenna, and Averroes on Intellect, Davidson; Appearance and Reality, Hacker; Aristotle and Neoplatonism in Late Antiquity, Blumenthal; see **** for the complete list). As is evident from Table 2, holdings and checkout rates became more balanced – with a trend, even, for the non-ethics books to be held and off shelf in greater numbers. The exclusion of the SEP titles also eliminated the tilt toward philosophy of mind and language in the non-ethics list.
Since the total number of missing books falls from 446 to 118, the power of the analysis is substantially reduced. However, the estimated effect size increases – 8.5% missing for ethics vs. 5.7% for non-ethics, for an odds ratio of 1.48 to 1 – and the difference thus remains statistically significant (95% CI for difference 0.3%-5.2%, p = .03). Note, also, that on this reduced data set ethics books are no more likely to be simply overdue than non-ethics books (indeed the trend reverses, though with a wide CI interval and high p value: 3.6% vs. 4.3% as a percentage off shelf, two proportion test, 95% CI for difference -2.5%-1.2%, p = .46).
----------------------------------------------------------------------------------------------------
Insert Table 2 about here
----------------------------------------------------------------------------------------------------
Use in law.
In conversation, several philosophers have suggested that the effect
might be driven largely by law students and professors, who presumably borrow
ethics books more often than non-ethics books in philosophy and may have
different library habits. The data do
not bear out this hypothesis. In the
Failure of independence? The two proportion tests I have been using assume the trials to be statistically independent. The present data might fail to be independent in two different and cross-cutting ways: Some titles might be more likely than others to go missing. (Consider Abbie Hoffman’s [1971] counterculture classic Steal This Book [not included in this survey].) Also, some libraries might be more likely to have missing books – if, for example, their security is lax or if a single patron is appropriating many books. (By the way, 13 copies of Steal This Book are held in the selected libraries. Of the four off shelf, three are missing.) Unfortunately, standard statistical tests for independence cannot be applied here due to the small number of relevant events for each title and for the smaller libraries.
However, we can examine the data by title and by library, to see if just a few titles or just a few libraries are driving the effect. If not, it may be reasonable to conclude that the independence assumption is approximately satisfied. I calculated missing as a percentage of off shelf for all titles with at least five books off shelf across all the sampled libraries, excluding the pre-1985 and the SEP titles (but not excluding law titles). Ethics titles are more likely to be missing, with a median of 8.3% for ethics vs 0.0% for non-ethics (Mann-Whitney, adj. for ties, p = .046). The difference is also statistically significant looking at the full set of titles (including pre-1985 and SEP titles), again excluding those with four or fewer off shelf (median 7.8% vs. 0.0%, p = .03).
Given a sample of only 32 libraries, many of which have only one or a few books missing, library-by-library comparisons will only be able statistically to detect large or consistent effects – especially since the data turn out to be non-parametric (with seven libraries missing only ethics books and one missing only non-ethics books). Looking library by library, on the reduced list (excluding SEP and pre-1985 titles but including law titles), 12 libraries have more ethics books missing than non-ethics books and 6 have more non-ethics books missing (the remaining libraries either have one of each missing or none) – a non-significant trend (binomial test, p = .14). Looking at the full list of titles, 25 libraries have more ethics books missing than non-ethics books and 2 have more non-ethics books missing. However, since there are twice as many ethics books checked out, overall, as non-ethics books on the full list, we should expect on average twice as many missing ethics books and non-ethics books in each library. A Wilcoxon ranked sign test of the proportion per library of (ethics books missing) / (ethics books missing + 2.0 x non-ethics books missing) vs. .5 is marginally significant (p = .07). Sixteen libraries are missing a greater proportion of ethics books than expected, while nine are missing a greater proportion of non-ethics books.
No particular topical subgroup of titles stands out post-hoc as particularly more or less likely to be missing than any other – not applied ethics, nor feminism, nor meta-ethics, nor books with a demanding as opposed to a lenient ethical view. However, there are too few titles in most of these categories, too little power, and too many confounds to test this systematically.
Discussion
The results so far suggest that ethics books are actually more likely to be missing than non-ethics books in philosophy. Furthermore, the estimated effect size increases when the relatively better-known books are excluded from the analysis, leaving only books likely to be checked out by professors and advanced students in philosophy. It does not appear that the results are driven by just a few titles or just a few libraries. However, given the potential non-independence of the trials and the statistical marginality of the library-by-library analysis, the conclusion can only be tentative. Therefore, I decided to do a second study examining prominent pre-20th-century philosophy texts. Are classic texts in ethics more or less likely to be missing than other classic texts in philosophy?
Since undergraduates likely check out a high proportion of classic titles, Study 2 is less pure a measure of the moral behavior of professional ethicists. However, it seemed advisable to confirm the trends of Study 1 on a different population of books, and it’s not unreasonable to suppose (looking at my own shelves and recalling those of others) that a substantial proportion of classic text checkouts are by professors and advanced students of philosophy. We cannot afford to purchase every edition of every classic text.
Study 2: Rates at Which Classic Philosophy Texts are Missing
Method
Compiling the list of titles. I turned again to the Stanford Encyclopedia of Philosophy (October 2006 edition). The preliminary lists were composed of texts before 1900 that appeared in at least five ethics bibliographies or five mind-and/or-language bibliographies. I removed titles prominent both in and outside of ethics (such as Hume’s Treatise) and that normally appear anthologized with books of the other type (such as Hume’s Enquiries and most of Plato’s dialogues). To flesh out the very short non-ethics list and to compensate for the mind and language bias in that list, I perused the historical entries of the SEP and added Bacon’s New Organon and Kant’s Critique of Judgment. All English and original-language editions were included; so also were anthologies in which the book in question appeared as the first-listed or most prominent work, unless the anthology included both ethics and non-ethics texts. I also made a third list of Nietzsche’s most prominent works. Although Nietzsche is in some sense an ethicist (as well as many other things), he condemns conventional morality. I thought it might be interesting to compare the rate at which his books are missing to the rate at which more traditional ethical works are missing. The final list of books appears in Table 4.
The lending libraries. I used the same
Collecting the data. A research assistant and I collected the data in late December, 2006, and early January, 2007, when most schools were between sessions, and then again in late February and early March, 2007, when all schools were in session. The Nietzsche data were collected only in March. Since undergraduates presumably borrow these books at high rates, I thought it might be informative to compare between-session and mid-session data. Books were again coded as on shelf, within period, overdue, or missing. Multi-volume works were coded separately if the library had a separate status line for each volume, unless the work was split into more than three volumes. (Some editions of Aquinas’s Summa, for example, are divided into many small volumes.) When volume data were combined, the highest coding number of all the individual volumes was used, with on shelf being 1, within period 2, overdue 3, and missing 4.
Results
As expected, a greater proportion of books was checked out in February-March than in December-January, though the effect was not large: 13.5% of ethics holdings and 15.9% of non-ethics holdings were off shelf during break, while mid-session the percentages were 16.7% (two proportion test, p < .001) and 18.4% (p = .02) respectively. As is evident from these percentages, non-ethics books were slightly more likely to be off shelf than ethics books, in contrast to Study 1. This difference was statistically significant over winter break (p = .01) and marginally significant mid-session (p = .08). Similarly to Study 1 (before the data reductions) ethics holdings (5541 mid-session) were about twice that of non-ethics holdings (2324 mid-session).
Overall, the mid-break and mid-session data look very similar, apart from the slightly higher off shelf rates mid-session and more copies of Plato’s Republic found mid-session (1119 vs. 884, probably due to different search strategies). The remaining analysis will treat the mid-session data only, since the Nietzsche data were only collected mid-session, and since the missing rates for the Republic are substantially lower mid-session, bringing them closer to the mean and thus erring on the side of statistical caution.
----------------------------------------------------------------------------------------------------
Insert Table 3 about here
----------------------------------------------------------------------------------------------------
As is evident from Table 3, ethics books are about twice as likely to be missing in this sample as non-ethics books, whether measured as a percentage of holdings (2.8% vs. 1.5%, 95% CI for difference 0.6%-1.9%, p < .001) or as a percentage of books off shelf (16.7% vs. 8.4%, 95% CI for difference 4.7%-11.9%, p < .001). The Nietzsche books are also more likely to be missing than the non-ethics books (5.6% vs. 1.5% as a percentage of holdings, 95% CI for difference 2.6%-5.5%, p < .001; and 19.5% vs. 8.4% as a percentage of off shelf, 95% CI for difference 5.8%-16.2%, p < .001). Measured as a percentage of holdings, the Nietzsche books are more likely to be missing than the ethics books (95% CI for difference, 1.3%-4.3%, p < .001), but by the more important measure of percentage missing among those off shelf, the difference is not statistically detectable (95% CI for difference, -2.4%-7.8%, p = .30). Though the numbers are small, overdue as a percentage of off shelf and of holdings is fairly similar between the three groups.
All libraries have at least twice as many ethics as non-ethics books missing – but there are also twice as many ethics books off shelf within period. As in Study 1, this was corrected by comparing the proportion, by library, (ethics books missing) / (ethics books missing + 2.0 x non-ethics books missing) vs. .5. The difference is statistically significant (t test, M = .71 vs. .5, 95% CI for mean .63-.78, p < .001). Twelve of the 13 libraries are missing a greater proportion of ethics books than expected.
By title the differences are also striking. Table 4 ranks titles by missing as a percentage of off shelf, with titles with fewer than 25 total off shelf in brackets. With the exception of Frege’s Foundations of Arithmetic, every single one of the 12 most-missing titles is either an ethics or a Nietzsche book. Likewise, of the eight titles least missing, only one is an ethics title – and that one is really more in philosophy of religion than in mainstream philosophical ethics. The mean percentage missing is 16.7% for ethics titles (not including Nietzsche) vs. 7.9% for non-ethics titles (odds ratio 2.1 to 1). Even with the tiny sample size (12 ethics titles, 9 non-ethics), this difference in means is statistically significant (t test, 95% CI for difference 2.1%-15.7%, p = .01). The three Nietzsche titles are missing at about the same rate at the ethics titles (mean percentage missing 19.0%).
----------------------------------------------------------------------------------------------------
Insert Table 4 about here
----------------------------------------------------------------------------------------------------
The lists can be made more comparable by cutting the outliers: books with fewer than 25 off shelf (see Table 4) and those with more than 100 (Aristotle, Hobbes, and Plato). The eight remaining ethics books have an average of 402 held, 62.5 off shelf, and 11.4 missing. The six remaining non-ethics books have an average of 329 held, 61.5 off shelf, and 4.8 missing. Missing as a percentage of off shelf on this reduced list is 18.2% for ethics and 7.9% for non-ethics, for an odds ratio of 2.3 to 1 (two proportion test, p < .001). The difference in mean, title-by-title, of missing as a percentage of off shelf – 18.2% vs. 7.1% (odds ratio 2.5 to 1) – remains significant (t test, 95% CI for difference 3.4%-18.7%, p = .01). Although the ethics titles still have higher holdings per book, the relationship between holdings and missing as a percentage of off shelf is relatively small and statistically undetectable. (On the full list of ethics titles r = .18, p = .53; for non-ethics titles r = .09, p = .80.) The difference in means remains if missing is calculated as a percentage of holdings (mean 2.6% vs. 1.3%, odds ratio 2.0, 95% CI for difference 0.2%-2.3%, p = .02).
Discussion
The results of Study 2 support the conclusion of Study 1. Looking at this very different selection of books, ethics books are again more likely to be missing. The effect is so large and consistent that it shows up even in low power title-by-title and library-by-library analyses.
Unlike Study 1, the effect may plausibly be driven mainly by undergraduates. However, the small difference between mid-term and between-term checkout rates suggests that few patrons are checking these books out by mid-term and returning them at term’s end, which one might expect to be a typical checkout pattern if undergraduates are borrowing the books for use in classes.
General Discussion
These data suggest that ethics books are more likely to be missing from academic libraries than other types of philosophy books. This effect appears to hold both for obscure books likely to be checked out mostly by professional philosophers and for widely read classics like Mill’s On Liberty and Descartes’s Meditations. If these data are representative, a philosophy book not on the shelf is anywhere from 25% to 150% more likely to be missing if it is an ethics books than if it is not.
The effect does not appear to be attributable to law faculty and students, since law books and ethics books in law libraries seem to be missing at about the same rate as other ethics books. Nor does it seem that ethicists are especially more absent-minded than non-ethicists, since the data on simply overdue books are mixed. The effects show up whether the data are analyzed by simple proportion test, by title, or by library. Furthermore, it does not appear that books endorsing conventional morality are less likely to be missing than books challenging conventional morality: Kant and Mill are missing as much as Nietzsche, Rawls as much as Williams.
With respect to their treatment of library books, then, it does not appear that the people reading philosophical ethics behave any better than those reading other sorts of philosophy; indeed, the opposite seems to be the case. It surely does not follow that ethicists and their students generally behave the same as, or worse than, people interested in other areas of philosophy. To draw this general conclusion would require more studies examining a diverse range of moral behavior. However, in one domain in which ethicists could have displayed superior conscientiousness, honesty, and concern for others’ property, they failed to do so.
There are potential confounds this study cannot control. Readers might more dearly love ethics books than other philosophy books. Readers of ethics books might be poorer than readers of other philosophy books and so more tempted to theft; or they might be wealthier and so more willing to risk fines. Ethics books may take longer to read and so be more likely to leave campus; or they may be more pleasant to read and so more exposed to the hazards of the cafe, the beach, and the bedstand. They are assigned for different sorts of course. They may have been more popular ten years ago than they are now. They may be more likely reported missing if a patron can’t find them on the shelf. A patron’s friends and spouses may be more likely to borrow them. Ethicists and their students might be busier than non-ethicists and their students. Thus the present data are at best only suggestive. Yet the fact that the data run consistently in the opposite direction of the effect that might have been expected implies that such confounding factors must be large if ethicists are indeed generally more conscientious with their library books than non-ethicists.
Those attracted to the view that philosophical moral reflection is morally useless or even harmful may find these data unsurprising. I cannot say that I myself am attracted to such a view. However, if in the end a broad range of research suggests that ethicists do not behave better than non-ethicists, we will have to reconsider the presumed role of reflection, or at least philosophical moral reflection, in generating moral behavior.
References
Aristotle
(4th c. BCE/1962). Nicomachean ethics. Trans. M. Ostwald.
Baier, C.J., & Wright, B.R.E. (2001). “If you love me, keep my commandments”: A meta-analysis of the effect of religion on crime. Journal of Research in Crime and Delinquency, 38, 3-21.
Blasi, A. (1980). Bridging moral cognition and moral action: A critical review of the literature. Psychological Bulletin, 88, 1-45.
Boyd, D.P. (1981). Improving ethical awareness through the business and society course. Business & Society, 20, 27-31.
Colby,
A., & Damon, W. (1992). Some do care.
Conroy, S.J., & Emerson, T.L.N. (2004). Business ethics and religion: Religiosity as a predictor of ethical awareness among students. Journal of Business Ethics, 50, 383-396.
Duizend, J., & McCann, G.C., (1998). Do collegiate business students show a propensity to engage in illegal business practices? Journal of Business Ethics, 17, 229-238.
Eckles, R.E., Meslin, E.M., Gaffney, M., & Helft, P.R. (2005). Medical ethics education: Where are we? Where should we be going? A review. Academic Medicine, 80, 1143-1152.
Eshuys, D., & Smallbone S. (2006). Religious affiliations among adult sexual offenders. Sexual Abuse, 18, 279-288.
Glenn, J.R. (1992). Can a business and society course affect the ethical judgment of future managers? Journal of Business Ethics, 11, 217-223.
Greene, J., & Haidt, J. (2002). How (and where) does moral judgment work? Trends in Cognitive Sciences, 6, 517-523.
Gross,
M.L. (1997). Ethics and activism.
Haidt, J. (2001). The emotional dog and its rational tail: A social intuitionist approach to moral judgment. Psychological Review, 108, 814-834.
Hiltebeitel, K.M., & Jones, S.K. (1992). An assessment of ethics instruction in accounting education. Journal of Business Ethics, 11, 37-46.
Hirschi, T., & Stark, R. (1969). Hellfire and delinquency. Social Problems, 17, 202-213.
Kant,
Kohlberg,
L. (1984). The psychology of moral development.
Krebs,
D.L., &
Loe, T.W., & Weeks, W.A. (2000). An experimental investigation of efforts to improve sales students’ moral reasoning. Journal of Personal Selling and Sales Management, 20, 243-251.
Luthar, H.K., & Karri, R. (2005). Exposure to ethics education and the perception of linkage between organizational ethical behavior and business outcomes. Journal of Business Ethics, 61, 353-368.
Martin, T.R. (2007). Do courses in ethics improve the ethical judgment of students? Business & Society, 20, 17-26.
Mill,
J.S. (1859/2003). On liberty. In M. Warnock
(Ed.), Utilitarianism and on liberty, 2nd
ed. (pp. 88-180).
Moody-Adams,
M.M. (1997). Fieldwork in familiar places.
Murphy, P.R., & Boatright, J.R. (1994). Assessing the effectiveness of instruction in business ethics: A longitudinal analysis. Journal of Education for Business, 69, 326-333.
Nussbaum,
M.C. (1997). Cultivating humanity.
Piaget,
J. (1932/1965). The moral judgment of the child.
Trans. M. Gabain.
Rest,
J.R. (1993). Research on moral judgment
in college students. In A. Garrod (Ed.),
Approaches to moral development (pp.
201-213).
Ross,
L., & Nisbett, R.E. (1991). The person and the situation.
Schwitzgebel, E. (2007). Still more data on the theft of ethics books. Blog post at The Splintered Mind (http://schwitzsplinters.blogspot.com), January 8.
Smith, P.L., & Oakley, E.F. (1996). The value of ethics education in business school curriculum. College Student Journal, 30, 274-283.
Stams, G.J., Brugman, D., Deković, M., van Rosmalen, L., van der Laan, P., & Gibbs, J.C. (2006). The moral judgment of juvenile delinquents: A meta-analysis. Journal of Abnormal Child Psychology, 34, 697-713.
Wynd, W.R., & Mager, J. (1989). The business and society course: Does it change student attitudes? Journal of Business Ethics, 8, 487-491.
Table 1
Study 1: all surveyed ethics and non-ethics books
|
|
Ethics |
Non-Ethics |
|
Holdings |
14,517 |
9,608 |
|
Off shelf |
3,721 |
1,775 |
|
Delinquent |
498 |
186 |
|
Missing |
323 |
123 |
|
Within period as % of holdings |
22.2% |
16.5% |
|
Delinquent |
||
|
as % of holdings |
3.4% |
1.9% |
|
as % of off shelf |
13.4% |
10.5% |
|
Overdue |
||
|
as % of holdings |
1.2% |
0.7% |
|
as % of off shelf |
4.7% |
3.5% |
|
Missing |
||
|
as % of holdings |
2.2% |
1.3% |
|
as % of off shelf |
8.7% |
6.9% |
Table 2
Study 1: older and more popular titles excluded
|
|
Ethics |
Non-Ethics |
|
Holdings |
5,040 |
5,601 |
|
Off shelf |
781 |
911 |
|
Delinquent |
94 |
91 |
|
Missing |
66 |
52 |
|
Within period as % of holdings |
13.6% |
14.5% |
|
Delinquent |
||
|
as % of holdings |
1.9% |
1.6% |
|
as % of off shelf |
12.0% |
10.0% |
|
Overdue |
||
|
as % of holdings |
0.6% |
0.7% |
|
as % of off shelf |
3.6% |
4.3% |
|
Missing |
||
|
as % of holdings |
1.3% |
0.9% |
|
as % of off shelf |
8.5% |
5.7% |
Table 3
Study 2: classic texts
|
|
Ethics |
Non-Ethics |
Nietzsche |
|
Holdings |
5,541 |
2,324 |
1,035 |
|
Off shelf |
926 |
427 |
298 |
|
Delinquent |
191 |
53 |
63 |
|
Missing |
155 |
36 |
58 |
|
Within period as % of holdings |
13.3% |
16.1% |
22.7% |
|
Delinquent |
|
||
|
as % of holdings |
3.4% |
2.3% |
6.1% |
|
as % of off shelf |
20.6% |
12.4% |
21.1% |
|
Overdue |
|
||
|
as % of holdings |
0.6% |
0.7% |
0.5% |
|
as % of off shelf |
3.9% |
4.0% |
1.7% |
|
Missing |
|
||
|
as % of holdings |
2.8% |
1.5% |
5.6% |
|
as % of off shelf |
16.7% |
8.4% |
19.5% |
Table 4
Study 2: Classics texts, by title
|
Author, Title |
Missing as % of off-shelf |
Ethics / Non / Nietzsche |
|
[Brentano, Psychology from an Empirical Standpoint] |
0.0% |
Non |
|
[Frege, Translations from the Philosophical Writings] |
0.0% |
Non |
|
Bacon, New Organon |
3.4% |
Non |
|
Kant, Critique of Judgment |
3.8% |
Non |
|
Descartes, Meditations |
5.6% |
Non |
|
Aquinas, Summa Theologica |
5.9% |
Ethics |
|
James, Principles of Psychology |
7.1% |
Non |
|
[ |
9.5% |
Non |
|
[Kant, Metaphysics of Morals] |
10.0% |
Ethics |
|
Kant, Critique of Practical Reason |
10.3% |
Ethics |
|
Kant, Critique of Pure Reason |
10.6% |
Non |
|
Bentham, Principles of Morals and Legislation |
11.1% |
Ethics |
|
Locke, Essay Concerning Human Understanding |
12.1% |
Non |
|
Aristotle, Nicomachean Ethics |
14.8% |
Ethics |
|
Hobbes, Leviathan |
15.0% |
Ethics |
|
Plato, Republic |
15.8% |
Ethics |
|
Nietzsche, Thus Spoke Zarathustra |
15.9% |
Nietzsche |
|
Kant, Groundwork of the Metaphysics of Morals |
16.2% |
Ethics |
|
Nietzsche, Beyond Good and Evil |
18.9% |
Nietzsche |
|
Nietzsche, On the Genealogy of Morals |
21.8% |
Nietzsche |
|
Mill, On |
23.2% |
Ethics |
|
Rousseau, Social Contract |
23.5% |
Ethics |
|
Locke, Two Treatises of Government / Second Treatise |
23.8% |
Ethics |
|
[Frege: Foundations of Arithmetic] |
26.3% |
Non |
|
Mill, Utilitarianism |
31.3% |
Ethics |