
Title: American Reference Books Annual (ARBA) Online
Publisher: Libraries Unlimited
URL: http://www.ARBAonline.com
Cost: $150-$350/year depending on size and type of library
Tested: March 6-9, 2007
Ready-reference sources are closest to my heart, as I grew up reading or at least extensively browsing (not just consulting) encyclopedias, dictionaries, factbooks and lexicons. They are also the most critical resources for reference librarians. In those rare moments when there is still money for buying or licensing ready-reference sources, it is of particular importance to learn what alternatives are available in the given reference genre for the disciplines and sub-disciplines, and how those reference sources were judged by reviewers.
This makes journals with very large review sections, such as Library Journal and School Library Journal, and the publications dedicated to book reviews, such as ARBA, Kirkus Reviews, Booklist, CHOICE, and Publishers Weekly so important for librarians.
That is the reason that I reviewed the last three in this column in the past year alone (and earlier, along with other important review collections, such as the New York Times Book Reviews, The London Review of Books and several other media review collections).
That is the reason why Thomson Gale, the host of this site, has been building for more than a decade this open access review collection which you are reading now, starting with James Rettig's outstanding critiques, which needed four of us (Blanche Woolls, David Loertscher, John Lawrence and myself) to resume when Jim accepted a job that required all his time.
ARBAonline cannot be directly compared with the other review collections as it is a special one, focusing on reference publications, whereas all of the others have a much broader scope, including among others, reference books. On the other hand, CHOICE has the academic libraries in its focus, School Library Journal serves the needs of school libraries and media centers, while Booklist and the other review journals focus on publishing reviews of books primarily relevant for the public libraries. ARBAonline covers ready reference books for all types of libraries, including the variety of special libraries. With this in mind, ARBAonline undoubtedly offers the largest collection of reviews ready reference sources in print and digital formats, for all types of libraries.
ARBAonline has nearly 17,000 reviews (by the time the mid-March, 2007 update will kick in). Almost half of the reviews are for books in Social Sciences. Humanities (including Arts) is the second largest segment, with about 4,650 reviews, and Science & Technology books have more than 3,000 reviews. More than 720 reviews are for books that belong to the General Reference category.
Unusually, the numbers on the FactSheet pages report lower numbers than I found through my tests. Typically, it is the other way around, the PR materials promising more than the database delivers. It reports to have 14,000+ reviews with 4,500+ titles in the Social Sciences, 3,000 in the Arts & Humanities; 1,600 in the Science & Technology field, and 400+ in the General Reference category.
As for the genres, encyclopedias and dictionaries lead the pack. I found reviews for 2,193 books with the word encyclopedia or encyclopaedia in the title, and for 1,753 books with the word dictionary in the title. For comparison, Choice had reviews for 1,537 and 1,065, search for the same query for the same time span as ARBAonline. The entire digital Choice collection had 2,283 and 1,065 records matching these search terms. Booklist Online (with coverage from 1992 to March, 2007), yielded less than 1,000 hits for the same title queries for the two genres, while Library Journal (for the period 1987-2007) produced an impressive total of more than 1,300, and School Library Journal checked in with 426 reviews for the two genres combined
There are reviews for more encyclopedias and dictionaries in all the review collections, because there are hundreds of books reviewed which don't have the word encyclopedia or dictionary or any of their variants in the title, but they are encyclopedias or dictionaries, such as most of those books which have A to Z in their title. Subject searches also would prove this, but the various review collections use different controlled vocabularies, thus a subject heading test search would not offer level playing field for the comparison.
The test showed that the largest publisher of ready reference books (and the host of this column), Thomson Gale (including its imprints) gets the most reviews, nearly 1,700, representing 10% of the entire ARBAonline collection. It is followed, not surprisingly, by the Greenwood Press Group (including Libraries Unlimited) with 1,611 reference books reviewed.
Some of the other largest publishers of print and digital ready reference books give a good sense of the scope of the reviews by publishers. McFarlands has 581, the Taylor & Francis Group (including Routledge and its other imprints) has 421, my favorite ready reference publisher, Dorling-Kindersley (to be searched as DK) has 224, Random House 199, Wiley 181, Oryx 165, and McGraw Hill has 149. There are reviews for more than 1,000 books published by various university presses combined.
The time span of coverage also may need explanation. It starts with the print equivalent of ARBA 1997, which in turn – in line with its policy- includes reviews of books published in 1995 (380), 1996 (1,528) and 1997 (1,750), so the actual years of coverage is more like 12 years than 10 years. Of course, I would like to have all the reviews of the print ARBA published since 1970, but I am not too much concerned about the 12-year time span even though the digital versions of Choice, and Library Journal go back to 1987, and Booklist Online's coverage starts in 1992.
Good reference books are regularly updated and published again and again, so it is likely that an excellent encyclopedia published in 1993 and reviewed in the print ARBA volume in the same year or the next (thus not covered by ARBAonline) had another edition a few years later, and another review, included in ARBAonline.
For example, the classic reference work, Magazine for Libraries, was first published almost 40 years ago, then updated and published time and again. ARBAonline has reviews for the 1997 and the 2000 editions. There are earlier and more current editions of this essential tool, but the two reviews are sufficient for an acquisition decision. There might be exceptions, especially in the areas of Arts and Humanities, but not too many.
As a compromise, to increase its time span with relevant records, Libraries Unlimited might consider adding reviews from the pre-1997 print ARBA volumes for award winning reference sources, such as the most recent editions of the ready reference masterpieces of Ken Kister from the mid 1990s about encyclopedias, dictionaries and atlases. These still provide excellent background information about evaluation criteria, and an intellectually rewarding reading experience, not just ready reference look-up purposes.
The reviews, which are written by librarians and staff members of Libraries Unlimited are not only signed, but also often include the reviewers' affiliation. The reviews are informative, critical to the necessary extent, and often somewhat longer than the reviews in Choice, and Booklist.
The matching terms are highlighted in the text in transparent yellow. Often the reviews provide good context by referring to other, competitive sources. If those were reviewed for ARBA, then their identifiers typically appear in the review. If those sources are included in ARBAonline, there is a link to them. I found some exceptions in my tests. For example, while the review of Historical Atlas of Islam, published by Harvard University Press in 2004, mentions and links to the reviews of The Atlas of Medieval Europe and to The Atlas of the Crusades, it does not mention, let alone links to the reviews in ARBAonline of the even more closely related two books published two years earlier by Brill Academic, and Continuum Publishing.
I found one perplexing link in the review of the Dictionary of Psychology of Andrew Colman (whose name is misspelled in the body of the review) published by Oxford University Press in 2001 when I wanted to check if there is a reference to the review of the better-known Dictionary of Psychology of Corsini published in 1999 by Taylor & Francis. There is no link to, neither mention of the Corsini dictionary, but there is a puzzling link in the concluding statement about the book being reviewed: “The Dictionary of Psychology (see ARBA 2001, entry 1343) will be a useful addition” (as if it were an unnecessary self-link). It turns out to be a link to the review of The Dictionary of Biology It is likely to be a misplaced link which should have appeared in an earlier sentence in which that dictionary was mentioned. Such errors or commission and omission were very rare in my tests.
On the other hand, there were plenty of references to reviews of the same book in the collections of Choice, RUSQ, Library Journal, School Library Journal, Booklist and Reference Books Bulletin (RBB) with the date of issue of the review. Although you can't expect such references in all the reviews, and they are missing, for example, for the Colman dictionary mentioned above and reviewed by Library Journal and Booklist with their full text available on Amazon.
These references are great if you have the print journal. In case of reviews from Booklist/RBB the situation is far better even if you do not have a subscription to the rather expensive Booklist Online digital collection. Why? Because although ALA does not offer the thousands of very good book reviews for free anymore, it apparently made a deal with Amazon where its full reviews (not just snippets) are posted.Understandably, one doesn't learn about this from the new, subscription-based Booklist website.
This means that you can quickly and freely look up most of the Booklist/RBB reviews in their entirety, as in the case of the Brain Encyclopedia whose review in Booklist well complements the review of ARBAonline. This would compensate many librarians for the ill-conceived and ill-handled elimination of the substantial, formerly open access review collection of Booklist on the ALA website that I lamented about in 2004 when the new ALA website was launched. Some of the full reviews from LJ, and SLJ are also available through Amazon for free. For reviews in Choice and ARBA I saw only free snippet at Amazon, which still might be useful.
Two years ago, when I looked at the first release of ARBAonline, I had problems with the software, and gave up the idea of reviewing it, in spite of the important content of the database. When I looked at the recently released new version, it had an immediate appeal with its clean design, swift search and a number of other features.
The software allows searching by author/editor, title, publisher, subject, keywords, publication year or year range, ARBA volume and publication date, and by broad subject categories and subcategories (referred to as content area filters).
The beauty of this latter feature is that it is not merely a search filter to refine the original search by title, author, keyword and or publisher, but also a direct search element. This means that you can practically browse a smaller subset, such as the 173 reviews of Women's Studies just by clicking on that area filter, and quickly sort them by publication year, publisher, author as dictated by your needs based on scanning the titles.
ARBAonline has beyond the usual Boolean and truncation options several others, including proximity and positional, and even frequency operators (which are not really Boolean operators, but this is just a nag). To the best of my knowledge, only Lexis-Nexis and Ovid offer the functional equivalent of frequency operator, which is a perfect antidote to the limitation of the Boolean yes/no, search term is present/absent world. It is especially useful in full text databases, but can come handy even in abstracting and review databases.
It allows the user to limit the search to items where the search term occurs more than n times. It can make the search more and more focused. Searching for the variants of the root word terroris* would find nearly 200 reviews. But using the command keyword (terroris*) >5 will reduce it to 24 hits, and raising the frequency minimum to an extreme value of 20, will still retrieve 7 hits.
Obviously, the higher the frequency of the search term, the more likely it is that the review not just mentions in passing the word terrorism, terrorist or some of the other variants, but it is indeed about a book focusing on the topic, and implies a substantial review of it, as you can see from a sample where the search term occurs more than 20 times. This precious feature (often used behind the scenes for relevance ranking) is also available as a specific and explicit sort options among other alternatives.
Records can be marked and saved (but only temporarily). I would like to see an option for downloading selected records in their entirety, or some of their elements as chosen by the user. It would be also useful to offer a link from the records to the library selected by the patron to see if the book is available, and to Amazon to see what other editorial reviews and other additional content and options are in that splendid digital bookshop, such as table of contents, excerpts, reviews, snippets of reviews, and/or the search inside the book, and/or the cited/citing references feature as I illustrated in my last review of Amazon. These can be done also through smart bookmarklets of short Javasripts, but building these into ARBAonline would make it even more attractive.
This new release of the digital version of ARBA shows what a difference a very good software can make for a time-honored valuable content, bringing the best out of it. The very reasonable yearly subscription fee (which depends on the size and type of the library and its patron population), make ARBAonline affordable even for small libraries at the $150/year subscription rate.