132 College & Research Libraries March 1999

Book Reviews As a Tool for Assessing Publisher Reputation

Matthew L. Jordy, Eileen L. McGrath, and John B. Rutledge

This article reports on the authors’ efforts to develop a method of using book reviews to establish the reputations of publishers. The authors ex­ amined the quality of books published by de Gruyter, Greenwood, Doubleday, University of Georgia Press, and Louisiana State University Press as it is expressed in abstracts of book reviews published in the online version of Book Review Digest. The authors extracted a sample for each publisher from Book Review Digest, examined the sample, and compared each publisher sample against a control sample. Although it is true that most book reviews are positive, there are discernible varia­ tions in how reviewers express themselves about books. The study also looks at Choice as a source of book reviews, and briefly examines the relationship between price and quality. This study adds to the literature of the use of book reviews as a selection tool.

“And as for the publishers, it is they collection development policy statements who build the fleet, plan the voy­ place the reputation of the publisher high age, and sail on, facing wreck, till on the list.1 The ALA itself has issued a they find every possible harbor that publication entitled Evaluating Informa­ will value their burden.”—Clarence tion: A Basic Checklist that asks the ques­ S. Day, The Story of the Yale Univer­ tion: What is the reputation of the pub­ sity Press Told by a Friend (1920). lisher, producer, or distributor?2 Specialized studies of selection method­ “Now Barrabas was a publisher.”— ology also recommend the reputation of Usually attributed to Byron. the publisher as a consideration. A study by John B. Rutledge and Luke Swindler ibrarians like publishers—ex­ cite “distinguished publisher” as a pri­ cept when it comes time to pay mary bibliographical consideration the bill. Librarians use the among other criteria for the selection of reputation of the publisher as monographs.3 If a theorem can achieve a prominent criterion in the selection of creedal status in librarianship, surely this books. Indeed, selection criteria found in one has.

Matthew L. Jordy, former Technical Support Manager in Davis Library at the University of North Caro­ lina at Chapel Hill, is a PDVC Lab Intern at Cisco Systems; e-mail: [email protected]. Eileen L. McGrath is the Collection Management Librarian in the North Carolina Collection at the University of North Carolina at Chapel Hill; e-mail: [email protected]. John B. Rutledge is the Bibliographer of West European Resources in Davis Library at the University of North Carolina at Chapel Hill; e-mail: [email protected]. 132 Book Reviews as a Tool for Assessing Publisher Reputation 133

The Quest for Quality How Reputation Is Formed Librarians affirm the importance of the If the ALA justly and necessarily approves publisher’s reputation because they know using reputation as a selection category, how much the publisher can add to the how do book selectors go about forming quality of a published book, from the ini­ an impression of a publisher’s reputation? tial selection of manuscripts to the distri­ Estimates of reputation can come from the bution for external review, the provision personal experience of research in a spe­ of editorial suggestions, and copyediting. cific discipline. The authors concur with A conscientious editor can significantly Paul Metz and John Stemmer that “most improve a manuscript in many ways.4 bibliographers’ impressions of most pub­ lishers represent an amalgam of conscious But are there too many publishers conclusions and much more visceral im­ for even a subject specialist to know pressions that have been gathered over them all through direct or personal years of academic training, personal read­ experience? ing, discussions with academic faculty and other librarians, inspection of library The reputation of the publisher serves receipts, and use of book reviews.”5 Most as an indispensable shorthand in book bibliographers will know a small num­ selection. Rarely is there enough time to ber of publishing firms well, but a larger assess each monograph for quality or to number much less well. wait for reviews to appear. Indeed, the Because publisher reputation must be publisher’s name often provides the only relied on for book selection, selectors known quantity that selectors have to use should be familiar with a large number in making the decision. It is a necessary of publishers. But are there too many pub­ shorthand because selection book-in­ lishers for even a subject specialist to hand usually is not an option. Even book­ know them all through direct or personal in-hand selection might not work: librar­ experience? If this is the case (and the ians can ascertain that the book has a authors think it is), does this result in scholarly look to it, but assessing its rela­ purchases based on merely brand-name tive value for its field among hundreds recognition or, worse, ill-informed preju­ of competing titles is nearly impossible. dices? The reputation of each press prob­ Recent advances in Web technology now ably varies slightly among groups of allow a selector to check the table of con­ people, with scholars holding one view, tents or to read a summary, but this is publishers another, and librarians yet time-consuming. The author is, of course, another. Few librarians and still fewer another piece of information available to scholars actually have the opportunity to the selector, but the author’s name may examine hundreds of works by a single be completely unknown, as is the case press with the intent of forming an opin­ with most first-time authors. One also can ion as to the quality of the product. As search to see whether the author has pub­ Metz and Stemmer point out, there is a lished other monographs, but the process “paucity of information” about publisher soon becomes circular: with whom has he quality and reputation.6 or she published, and what reputations do those publishers have? A Trial Run Publisher reputation forms the basis The high prices charged by certain pub­ for some approval plans. One of the most lishers for their monographs strain our common types of plans is that designed ability to see publishers as partners in the to cover university presses; in effect, these educational process. Selectors must make plans give sanction to an entire class of decisions about expensive items with the publishers. But there are different “classes” same paucity of information, although of university presses. Reputations vary. Is increasingly there are timely electronic the quality consistently high? sources of information about book con­ 134 College & Research Libraries March 1999 tent. Still, it is time-consuming to pursue itself put certain limits on the investiga­ this information. With book prices for tion. For example, Book Review Digest in­ academic titles now averaging above cludes very few foreign-language titles. $57.85, do expensive monographs war­ Thus, this study confines itself to English- rant the cost?7 language titles. Initially, concerns about high prices for Many selectors make a mental list of monographs led the authors to choose the presses thought to be “questionable,” if publishing firm de Gruyter for study. not actually disreputable. It would have With de Gruyter titles costing an average been interesting to examine the quality of $158.89, this imprint begged for inves­ of certain controversial presses (which tigation. The authors first undertook a must remain nameless). Unfortunately, crude test to see if a method could be de­ Book Review Digest contained too few re­ veloped that looked promising. They ex­ views of their imprints to give a repre­ amined sixty book reviews of de Gruyter sentative and statistically valid sample. titles in the electronic version of Book Re­ This in itself is telling. view Digest, one of the FirstSearch group of databases.8 Reading sixty mainly posi­ Previous Work on Book Reviews tive reviews suggested that a de Gruyter A number of interesting library-focused product will very likely be of high qual­ studies of book reviews were identified ity. More important, examination of a before the study method was completed. manageable number of book reviews per­ Scholarly studies of book reviews have suaded the authors that developing a repeatedly raised issues that are troubling more rigorous method for using Book with regard to the application of book Review Digest as a tool to assess pub­ reviews for the purpose of book selection. lisher quality would lead to significant Important among them is the role of jour­ results. nal editors, who exercise a large measure of discretion (and power) in deciding Goals and Method which books are reviewed. There is never The trial run encouraged the notion of enough space to review all books pub­ developing a method for forming a rea­ lished. One cannot exclude the possibil­ sonable opinion about the quality of the ity that both book review editors and re­ books published by several presses. The viewers are subtly and unconsciously authors also wanted to test the hypoth­ influenced by the reputation of the pub­ esis that certain publishers consistently lisher. In all likelihood, many “bad books” produce high-quality monographs. Ide­ (however that might be defined) do not ally, the method would allow for a com­ make it past these gatekeepers. A sad cor­ parison of the reputation of various types ollary to this rule is that some good books of presses publishing books suitable for also are not reviewed because of inad­ research libraries. Finally, the authors equate marketing and promotion. wanted to examine the relationship be­ As mentioned earlier, book reviews tween their results and those of earlier tend to be positive. Judith Serebnick dis­ studies on book reviews. covered that “the great majority of re­ Book Review Digest is truly a remark­ views are favorable.”9 “[M]ost reviews are able tool in that it allows users to gather found to be positive,” Dana Watson re­ many book reviews very quickly. The ported.10 Even Choice, a tool for the library authors looked carefully at the list of jour­ profession, recommended 75 percent of nals from which reviews are culled. With the books reviewed for purchase “with few exceptions, these journals fit into the few or no reservations.”11 In an editorial collecting profiles of larger academic li­ in Library Journal, Francine Fialkoff main­ braries. Books reviewed by these sources tained that 85 to 90 percent of reviews are are very likely to be acquired by academic positive.12 Worse, in recent years, “grade libraries. Unfortunately, however, the tool inflation” has infiltrated the world of Book Reviews as a Tool for Assessing Publisher Reputation 135 book reviews. Robert J. Greene and looked at those reviews where there were Charles D. Spornick found a decline in three different opinions and tried to reach unfavorable book reviews from nine to a common understanding; after that, they five percent from the late 1980s to the looked at cases where two agreed and one early 1990s.13 disagreed. Avoiding subjective evalua­ tions was not easy. The authors had to Developing Objective Categories, learn to focus on what the reviewer re­ Minimizing Subjectivity ally thought about the book, rather than Before settling on a workable ranking on their own opinions of it. Ironing out system to delineate quality, the authors the differences through conversation al­ experimented with various systems us­ lowed the authors to reach a common un­ ing several sample batches of reviews of derstanding of the categories and thus works by diverse presses. Despite the fact removed the arbitrariness that can creep that most books get positive reviews, it into individual assessments. was possible to make some valid and use­ ful distinctions. The authors devised a Method system that relies on broad distinctions All the presses selected for this study they think are intuitive. The first category publish books routinely purchased by (rated 1) was reserved for the outstand­ university libraries. To some extent, they ing book. Such exemplary titles receive are in competition for libraries’ scarce extremely positive or almost wholly lau­ dollars. Although the authors’ special in­ datory reviews. Often a descriptive ad­ terests influenced the selection of certain jective such as “outstanding” or “magis­ presses for the study, every attempt was terial” compels this categorization. made to apply the method to a range of Second, many books are perceived by re­ presses. First, the authors developed a viewers as very good (rated 2), even pool of publishers of professional inter­ though they may contain flaws in est to them. Next, they searched Book Re­ method, content, or style. However, more view Digest by publisher. When a pub­ books fall into the average, adequate, or lisher had a large number of titles in the “pretty good” category (rated 3). Such database, a representative sample was works attract praise and criticism in about taken. The number of reviews per pub­ equal measure but can still be recom­ lisher for the firms of interest to the au­ mended by the reviewer. A fourth cat­ thors ranged from a low of 99 for de egory was reserved for that small percent­ Gruyter to a high of 1,386 for Doubleday. age of books that receive a mostly Selecting eighty-one (in the case of de negative review (rated 4). Finally, a fifth Gruyter, eighty-six) reviews per publisher category (rated 0) includes those reviews produced a sample large enough to be that are chiefly descriptive in nature or representative of the whole. Individual do not provide enough information to reviews were extracted and printed out permit an assessment.14 (one review per page), in reverse chrono­ Book reviews can be subjective and are logical order as presented in the database. not immune from politics. Early on in the Next, the authors examined the sample research, the authors had to learn to read and discarded any duplicates. However, the reviews in a way that neutralized the they did not exclude reviews of the same subjectivity they brought to the process book by different reviewers. Each review as individuals. All three authors exam­ received an accession number for pur­ ined every review in each batch, and each poses of identification. Then the authors made an independent assessment mark­ read each review independently and ing 1, 2, 3, 4, or 0 on the reverse side of evaluated it using the established crite­ the text review printout. When the assess­ ria. With each batch, differences of opin­ ments were completed, the authors met ion had to be sorted out and the rankings to resolve their differences. They first harmonized. After any differences were 136 College & Research Libraries March 1999

views do not, in fact, serve as critiques TABLE 1* but, rather, simply announce the fact of Control Sample publication. Nearly half (42%) of the con­ trol batch fell into the broad third cat­ Category Count Percentage egory. 1 10 10% 2 22 22 University Presses

3 42 42 The university press has been defined as

4 10 10 “an organization whose function [i]s to

0 16 16 publish works which no one would

Totals: 100 100% read.”15 (Precisely the books that aca­ * Figures in some of the following demic libraries seek to acquire!) Because tables may not total 100% due to so many academic libraries routinely pur­ rounding chase university press titles, two univer­ resolved, the data were entered into a sity presses were included in the study. It spreadsheet. The ratings were stored in is a fact of economic life that many uni­ the spreadsheet along with the date of the versity presses have developed special­ review, the title of the book, the name of the ties in regional publications, reasoning publisher, the date of publication, and the that they know and serve their local mar­ journal in which the review first appeared. kets best. Natural curiosity and enlight­ ened self-interest led the authors to Control Sample choose two presses that have particular It soon became evident that an objective relevance for their own collections on the standard of comparison was needed to American South. The possibility that the put the assessment of the several types of results might contrast somewhat with presses in context. It was decided to draw those found for a large, international, aca­ a random sample of all the reviews in Book demic publisher such as de Gruyter made the Review Digest to serve as a control batch. choice of two university presses attractive. Using a table of random numbers, one The University of Georgia Press and hundred reviews were extracted to rep­ Louisiana State University Press have resent everything contained in the data­ fairly large annual title productions, but base without restriction by press. The con­ not nearly as large as the behemoths Ox­ trol batch allowed the authors to think ford and Cambridge, or Chicago, the larg­ about the typical or average “grade” est American university press.16 Yankee earned by books in Book Review Digest (see Book Peddler reports that the average table 1). price of a University of Georgia Press title The sample batch corroborated the is $35.55; Louisiana State’s books are rela­ findings of earlier studies that book re­ tively inexpensive at $26.14.17 Founded in views are overwhelmingly positive. Fully 1935, LSU Press’s output is a respectable 74 percent of the books in the sample re­ ceived a positive evaluation. What was TABLE 2 most surprising was that 10 percent of the books in the sample were considered out­ Table for University of standing by the reviewers. This was taken Georgia Press Reviews to be strong evidence of “grade inflation” in book reviews. At the opposite end of Category Count Percentage the scale, the same percentage (10%) of 1 4 4.94% books earned a mostly negative review. 2 35 43.21 Many of the reviews (16%) simply could 3 31 38.27 not be categorized because they concen­ 4 4 4.94 trated on summarizing the book’s content 0 7 8.64 rather than critiquing it. Many book re­ Totals: 81 100% Book Reviews as a Tool for Assessing Publisher Reputation 137

ceptional. The largest number of reviews TABLE 3 (thirty-five), or 43 percent, fell into the Table for Louisiana State second category, extremely positive. University Press Reviews Thirty-one reviews (38.3%) indicated that the book under consideration was of Category Count Percentage mixed quality, but still a worthwhile con­ 1 2 2.47% tribution. Only four reviews were mostly 2 24 29.63 negative (almost 5%).

3 43 53.09 The pattern for LSU Press titles re­

4 8 9.88 viewed differed slightly (see table 3).

0 4 4.94 Eighty-one of 509 reviews in the database

Totals: 81 100% were selected (retrieval date: 11/09/97). seventy titles per year with approxi­ Reviewers were clearer in stating their mately one thousand titles in print. The opinions about LSU books: only four of press has “a special emphasis on south­ the eighty-one reviews selected for LSU ern history and literature.”18 did not contain sufficient information to The University of Georgia Press was allow for an assessment of the reviewer’s established in 1938 and admitted to the opinion. The overwhelming majority of American Association of University Pub­ reviews were positive, but only two of the lishers in 1940. In the past decade, it has titles (2.47 percent) merited a designation published sixty to ninety new titles per of outstanding—half as many as Georgia. year of literary criticism, American his­ The percentage of LSU reviews that fell tory, Southern studies, and scholarly into categories 1 and 2 was significantly monographs in related fields. Like many less than comparable numbers for the university presses, it has begun to University of Georgia Press. Many more broaden its offerings through the inclu­ titles from LSU fell into the third category. sion of memoirs, literary titles, and popu­ In fact, two-thirds of all the reviews con­ lar works on its home state. The authors sidered the LSU books under review selected this press for the study because worthwhile, but flawed. Eight of LSU’s of their interest in Southern studies and reviews—twice the number for Georgia— because of a sense that it might become received a mostly negative assessment. as important a publisher in this field as The negative reviews came from a broad LSU Press or their local favorite, the Uni­ range of publications, from The Economist versity of North Carolina Press. to Booklist. Editors struggle to produce high-qual­ Two University Presses Compared ity books, and competition among uni­ Book Review Digest contained 429 reviews versity presses can be intense. Persons of University of Georgia Press publica­ involved in scholarly publication prob­ tions (retrieval date: 11/25/97). Eighty- ably will find these differences to be of one reviews were selected for examina­ tion. Seven of these abstracts (8.6%) did TABLE 4 not provide sufficient information to de­ Table for de Gruyter termine the reviewer’s estimation of the Reviews book’s quality. (These were either short reviews from Booklist and Choice or, sur­ Category Count Percentage prisingly, longer reviews from the New 1 13 15.12% York Times Book Review or the Times Liter­ 2 23 26.74 ary Supplement.) The overwhelming ma­ 3 40 46.51 jority of reviews (86.4%) were positive 4 7 8.14 (see table 2). Four of the eighty-one re­ 0 3 3.49 views (almost 5%) ranked the book un­ Totals: 86 100% der consideration as outstanding or ex­ 138 College & Research Libraries March 1999

A sample of eighty-six de Gruyter titles TABLE 5 taken from Book Review Digest yielded the Table for Greenwood results shown in table 4. Fifteen percent Reviews of the titles in the sample fall into the out­ standing category, a high percentage in­ Category Count Percentage deed! De Gruyter placed many more titles 1 5 6.17% in the outstanding category than did ei­ 2 27 33.33 ther of the two university presses; no 3 36 44.44 other publisher came close to this num­

4 11 13.58 ber. However, this high standard could

0 2 2.47 not be maintained in the second category,

Totals: 81 100% in which reviewers adjudged only about 27 percent of de Gruyter publications to greater significance than might appear at be very good. Although this is higher than first glance. If a large number of titles from the control batch, it is lower than the univer­ any university press fell into the fourth sity presses and the academic trade presses. category, serious questions about that Just under half of de Gruyter’s books press’s reputation and the quality of its (46.5%) merited only a middling ranking, list would have to be raised. Although the whereas about 8 percent earned mostly authors did not have time to examine a negative comments from reviewers. Why large number of university presses, it is do so many de Gruyter titles fall into the probable that reviews of their publishing third category rather than the second? In output would conform to one of these this regard, is de Gruyter an “ordinary patterns. Publications of the “first-rank” publisher?” How does one evaluate this university presses probably conform situation? Does de Gruyter lack sufficient more closely to the de Gruyter pattern. editorial staff to shape a larger percent­ Although this is admittedly speculation, age of its titles into outstanding or very the authors have provided a method that good works? What are the practical lim­ allows anyone to explore the reputation its to a drive for quality? Perhaps the over­ of a wide range of publishers within a all results have to do with the total vol­ short period of time. ume of books published, excellence be­ ing harder to achieve in great quantities. Return to de Gruyter Having refined the methodology, the au­ Greenwood thors reexamined more rigorously the Greenwood Press is one of the five im­ reputation of de Gruyter as it is revealed prints of the Greenwood Publishing by the reviews. The name de Gruyter re­ Group. Originally known as a reprint fers to a family of publishers: Walter de publisher, Greenwood Press now pro­ Gruyter, which took over Mouton (now duces reference works and scholarly Mouton de Gruyter) in 1977, and Aldine monographs in the humanities and the Publishing Company (now Aldine de behavioral and social sciences. Well Gruyter) in 1978. It is the yearly output known to academic librarians, Green­ of these three publishers—approximately wood was one of the publishers included 350 titles per year—that is under discus­ in the Metz and Stemmer study. The av­ sion here. De Gruyter describes itself as erage price of a Greenwood title, as re­ an “international academic publishing flected in Yankee’s approval plan cover­ house situated in Berlin,” publishing in age, is $62.90.20 almost all fields of knowledge, primarily Table 5 shows the results for Green­ in English. Its titles tend to stay in print wood. Just over 6 percent of Greenwood’s for long periods of time. Currently, the titles achieved a ranking of outstanding. number of de Gruyter titles in print is Reviewers generally express themselves about 12,000.19 positively about Greenwood titles. Con­ Book Reviews as a Tool for Assessing Publisher Reputation 139

Any editor could rejoice to see that about TABLE 6 74 percent of the firm’s publications were Table for Doubleday positively reviewed. However, about 15 Reviews percent drew heavy fire from the critics. This is the highest percentage of fourth- Category Count Percentage category books reported in the study and 1 2 2.47% higher than the control batch. 2 19 23.46

3 39 48.15 All Presses Viewed Together

4 12 14.81 When the various presses examined are

0 9 11.11 compared with the control batch (which

Totals: 81 100% represents the generality of books), the differences tend to show up most strik­ sequently, exactly one third (33.33%) of ingly at the extremes. Looking only at Greenwood’s titles were rated in the very books in the control batch that received a good (second) category, and 44.44 percent ranking of outstanding or very good (cat­ were rated to be of average quality (third egories one and two taken together), one category). Interestingly, more Greenwood finds that 32 percent of the books in the titles fall into the mostly negative cat­ control batch fell into the two upper cat­ egory (13.6%). What editorial factors or egories (see table 7). In this respect, economic considerations produce this Doubleday and LSU are typical, and kind of result? Is a lack of editorial over­ Greenwood, de Gruyter, and Georgia per­ sight the cause, or is there a multiplicity form much better. If the third and fourth of factors that could include such diverse categories are aggregated, Doubleday and matters as corporate profit margins or an LSU imprints have a higher concentration insufficiently critical customer base? at the bottom than do Greenwood, de Gruyter, and Georgia. Looking at the Doubleday fourth category alone, one notes that only Doubleday is an old, well-known com­ the University of Georgia Press was able mercial publisher of both popular and lit­ to avoid a significantly lower number erary fiction, as well as general interest here (4.94%) than the control sample nonfiction. It is now part of the Bantam (10%). Results for de Gruyter are compli­ Doubleday Dell Publishing Group, Inc., cated by the fact that most of the titles in which is itself part of the international the sample were in the social sciences. Social Bertelsmann AG. All the titles reviewed sciences titles tend to be reviewed slightly less in this study carried the Doubleday im­ favorably than humanities titles.21 print rather than mass-market Bantam or The authors would very much like to popular fiction Dell imprints. see Book Review Digest expanded to in­ On the whole, Doubleday titles receive clude a larger number of review sources good marks from reviewers (see table 6). and more foreign-language titles. Librar-

TABLE 7 Table of All Presses in Study

Rating Control deGruyter Doubleday Greenwood LSU UGA 1 10.00% 15.12% 2.47% 6.17% 2.47% 4.94% 2 22.00 26.74 23.46 33.33 29.63 43.21 3 42.00 46.51 48.15 44.44 53.09 38.27 4 10.00 8.14 14.81 13.58 9.88 4.94 0 16.00 3.49 11.11 2.47 4.94 8.64 Totals: 100.00 100.00 100.00 100.00 100.00 100.00 140 College & Research Libraries March 1999

content. Although editorial attention to TABLE 8 detail drives up production costs, Euro­ Table for Choice Reviews pean labor and distribution costs are other factors that must also be taken into ac­ Category Count Percentage count. The inelasticity of the market for

1 11 8.8% books would have to figure in the discus­

2 41 32.8 sion as well. If cost really did guarantee

3 57 45.6 quality, many more of de Gruyter’s titles

4 9 7.2 should have fallen into the second cat­

0 7 5.6 egory, yet three other presses (Green­

Totals: 125 100% wood, Georgia, and LSU) all “outscored” ians might benefit from greater inclusion de Gruyter in this category. The “edito­ of titles in the hard sciences where the rial manpower” at de Gruyter may not price of individual monographs can be be adequate to produce a larger number even more expensive. Publications by of its publications in the first category. some of the most expensive publishers in Likewise, is it a coincidence that the pub­ the world rarely appear in reviews cov­ lisher with the least expensive books, ered by Book Review Digest. LSU, had more books in the third category than the other presses? The lower aver­ Choice As a Source of Book Reviews age price may be due to the number of fic­ Because many of the reviews in the study tion titles in LSU’s list.22 Because de came from Choice, it was possible to es­ Gruyter publishes no fiction or general- tablish a profile of that journal as a source interest books, the higher cost for its books of book reviews. Table 8 shows the fig­ is for specialized scholarly materials. ures for Choice. It is interesting to com­ pare reviews in Choice with the reviews Librarian Assessment of Quality found in Book Review Digest as represented versus Reviewer Assessment by the control batch. Choice reviewers are Librarians queried by Metz and Stemmer not far from the norm in most respects. collectively gave Greenwood a 3.50 on a They are no more lavish with praise than five-point scale of perceived quality.23 the typical reviewer: 8.8 percent of books Doubleday received 3.00. Is it revealing in Choice were judged outstanding, com­ of librarians’ attitudes toward publishing pared to 10 percent in the control. How­ in general that the survey respondents ever, Choice reviewers proved signifi­ were unwilling to label many presses as cantly more generous in praising a book low in quality? (Only three presses in as very good (32.8%) than the typical re­ their list rated below 3.00—University viewer (22%). But even for Choice, most Press of America, Haworth, and Mellen.) books are no more than pretty good But librarians are likely right in their as­ (45.6%), just as they are for most review­ sessment. This is borne out by the reviews ers (42%). It should be noted that few re­ (most of which were surely written by views in Choice fail to state an opinion. nonlibrarians) in the study. Doubleday had the highest number of books in the Is There a Relationship between fourth category and a low number in the Price and Quality? top category. Reviews for Doubleday Is it a coincidence that the publisher with showed the highest percentage in the bot­ the highest percentage of outstanding tom two categories taken together. books (de Gruyter) is also the most ex­ Greenwood, at 3.50, is perceived by pensive publisher in the study? The au­ Metz and Stemmer’s librarians to be thors do not pretend to be economists, but slightly better than Doubleday (3.00). In the high cost of books from northern Eu­ the top two categories of the study taken rope is surely due to several factors, not together, Greenwood did better than just to high standards for production and Doubleday, better even than one univer­ Book Reviews as a Tool for Assessing Publisher Reputation 141 sity press. In this case, it seems that librar­ the part of librarians. The method pre­ ians ought to have ranked Greenwood sented in this study can provide that ca­ somewhat higher than they did in the pability. It is hoped that more librarians Metz and Stemmer survey. These mixed will use this method to evaluate the qual­ results (the librarians were right about ity of the output of publishers whose Doubleday, but not quite so accurate works they regularly purchase. Additional about Greenwood) point to the need for studies can lead to a more precise estimation greater familiarity with the presses and of publisher reputation, given its inevitable, more accurate assessment capabilities on and important, role in book selection.

Notes 1. It has become easier to examine collection development policies now that libraries are mounting them on the World Wide Web. Almost one quarter of ARL libraries have some form of collection development statement on the Web, as do non-ARL academic libraries such as Mansfield University, and many public libraries. 2. Evaluating Information: A Basic Checklist (Chicago: ALA, 1994). Not paginated. 3. John B. Rutledge and Luke Swindler, “The Selection Decision—Defining Criteria and Es­ tablishing Priorities,” College & Research Libraries 48 (Mar. 1987): 129. 4. For a warm testimonial on the work of a good editor, see Francis Paul Prucha, “Livia Appel and the Art of Copyediting,” Wisconsin Magazine of History 79 (summer 1996): 364–80. Unfortunately, the number of in-house editors has declined and some authors are now seeking editorial help from other quarters. See “As Publishing Pressures Rise, So Do Errors,” New York Times, June 29, 1998. 5. Paul Metz and John Stemmer, “A Reputational Study of Academic Publishers,” College & Research Libraries 57 (May 1996): 235. 6. Ibid., 235–36. 7. Trialogue: Publishing News for Publishers, Vendors, and Librarians, no. 7 (spring 1998): 9. Trialogue is published by Yankee Book Peddler, Inc. Average cost figures are based on Yankee’s U.S. approval plan coverage. The authors think that it thus accurately represents the average prices of books that academic libraries are likely to purchase. 8. Book Review Digest is one of the databases in OCLC’s FirstSearch service. It covers a hun­ dred periodicals, many of them scholarly and intellectual. The database contains approximately 403,000 records (as of August 28, 1998). 9. Judith Serebnick, “An Analysis of Publishers of Books Reviewed in Key Library Jour­ nals,” Library and Information Science Research 6 (July–Sept. 1984): 301–2. 10. Dana Watson, “Reviewing: A Strategic Service,” in A Service Profession A Service Commit­ ment: A Festschrift in Honor of Charles D. Patterson, ed. Connie Van Fleet and Danny P. Wallace (Metuchen, N.J.: Scarecrow Pr., 1992), 23. 11. Paula Wheeler Carlo and Allen Natowitz, “The Appearance of Praise in Choice Reviews of Outstanding and Favorably Assessed Books in American History, Geography, and Area Stud­ ies,” Collection Management 20 (1996): 102. 12. Francine Fialkoff, “Too Many Positive Reviews?: Librarians/Publishers/Book Review Edi­ tors Disagree on the Answer,” Library Journal 119 (Jan. 1994): 90. 13. Robert J. Greene and Charles D. Spornick, “Favorable and Unfavorable Book Reviews: A Quantitative Study,” Journal of Academic Librarianship 21 (Nov. 1995): 252. 14. The reviews in Book Review Digest sometimes do not reveal how the reviewer really feels about the book being reviewed. The authors wondered if the full versions found in the source journals might contain better information. Some comparisons of the full version of the review with the summary presented in Book Review Digest convinced the authors that the full version usually did not offer much more evaluative information than the abstract. Tracking down the full version of hundreds of book reviews would have been beyond reasonable time limits for this project. 15. Cited by Sheldon Meyer in “University Press Publishing,” in International Book Publishing: An Encyclopedia (New York: Garland, 1995), 355. 16. Ibid., 357. 17. Trialogue no. 7 (spring 1998): 6. 18. See “Bienvenue à LSU Press” at the LSU Press Web site: http://www.lsu.edu/guests/ lsuprss/welcome.html. Site visited July 14, 1998. 19. See the publisher’s own Web site at: http://www.degruyter.de/history.html. 142 College & Research Libraries March 1999

20. Trialogue no. 7 (spring 1998): 9. 21. Greene and Spornick, “Favorable and Unfavorable Book Reviews,” 252. 22. LSU Press has now discontinued regular publication of new fiction. See “Submissions” at: http://www.lsu.edu/guests/lsupress/index.html. Site visited July 14, 1998. 23. Metz and Stemmer, “A Reputational Study of Academic Publishers,” 238, 240.