Introduction. Article and book reading patterns of scholars: findings for publishers

Carol Tenopir, Rachel Volentine, and Donald W. King Article and book reading patterns of scholars: findings for publishers 279 Learned Publishing, 25: 279 291 doi:10.1087/20120407 Article and book reading patterns of scholars: findings for publishers Introduction Previous studies of the scholarly article reading patterns of academics show that there are differences in some reading behaviors depending on personal characteristics of readers (e.g. age, subject discipline, or work responsibilities) or on characteristics of the instances of reading (e.g. purpose of reading). 1 Scholarly book reading, however, has been less often studied. Knowing more about these differences and reading patterns should help publishers and librarians design more effective book and journal systems and services now and into the future. This paper examines differences in article and book reading patterns based on personal characteristics of academic staff (faculty members) in the United Kingdom. Six higher learning institutions participated in the 2011 survey, which was funded by JISC Collections and led by the University of Tennessee. The survey builds on reading surveys conducted by Carol Tenopir and Donald W. King in the US since 1977 2 and in Australia and Finland in 2005 and 2006. 3 The full report is available at http://www.jisc-collections.ac.uk/news/uk-scholarly-reading/. Research questions from the study that are explored in this article include: Do academics in different subject disciplines read differing numbers of articles and books? Do younger academics read a different number of articles and books than their older counterparts? Do reading patterns, such as source of article or book, time spent per reading, or format of article or book, differ by demographic characteristics of readers? Do academics who spend more time on research or teaching differ in their reading? Do high achievers, i.e. academics who publish more on average or who have won Article and book reading patterns of scholars: findings for publishers Carol TENOPIR, Rachel VOLENTINE, and Donald W. KING University of Tennessee ABSTRACT. Surveys of academic staff in six universities in the UK provide insights for publishers into scholarly article and book reading patterns of academics and differences based on personal characteristics of readers. These surveys were part of the 2011 UK Scholarly Reading and the Value of the Library Study funded by JISC Collections and based on studies conducted by Tenopir and King since 1977. Scholarly articles, especially those obtained from the library s e-journal collections, are a vital part of academic work. Reading patterns of books are quite different than articles; books most often come from personal print collections. Book readings are still important for research and teaching, however, especially for humanists. Academics come into contact with multiple sources of information every day and therefore, convenience and easy access are important factors. Knowing more about academic reading patterns helps publishers and librarians design more effective journal systems and services now and into the future. Part of a study done for JISC Collections, 2012. A full reportcanbefoundatcanbefoundat http://www.jisc-collections.ac.uk/news/uk-scholarly-reading/ Carol Tenopir, Rachel Volentine, and Donald W. King 2012 Carol Tenopir Rachel Volentine Donald W. King

280 Carol Tenopir, Rachel Volentine, and Donald W. King 2,117 respondents answered at least one question awards for teaching or research in the last two years, read more articles and books than others? Do frequent users or creators of social media read or publish less in traditional scholarly resources such as journals or book? Presented here are ten lessons for publishers based on the answers to these questions and other survey findings. Previous studies Tenopir and King 2 and King and Tenopir 4 summarize reading patterns of science and non-science academic staff members through the 1990s. These two sources provide extensive literature reviews and serve as background for the data presented in this report. Tenopir et al. 1 focus on factors influencing reading patterns of academics across subject disciplines in the US. Factors that influence reading patterns include subject discipline, work responsibilities, productivity of faculty, age of faculty members, and the purpose of reading. Other multi-university studies focus on how academics use electronic journals, online resources, and libraries. 5 Talja and Maula 6 examine the effects of subject discipline on search patterns. Other studies demonstrate that different disciplines have varying availability of electronic sources 7,8 and show that staff members in the sciences prefer and read more electronic journal articles than in humanities or social science disciplines. 9 A recent Research Information Network (RIN) study demonstrates that across the six disciplines of chemistry, environmental sciences, economics, life sciences, physics, and history, academic staff members in life sciences are the most likely and staff members in history are the least likely to use e-journals on a daily basis. 10 The RIN study also found that interviewees from all disciplines claimed that journal literature formed the bulk or the main body of literature consulted, for the science disciplines, journals were said to form around 95% (100% in some cases) of this with virtually all of it available (and used) electronically. 11 Recent studies have examined the role of e-books in academia. A report by CIBER 12 found that nearly two-thirds of teaching staff and students have used an e-book to support their work or study or for leisure purposes, and more than half of users said the last e-book they used was provided by their university library. Previous studies concentrate on either articles or books and typically on general behaviors, rather than details of scholarly article and book reading. Our study takes an in-depth look at academic staff reading patterns of both articles and books. Methodology and demographics of respondents As with previous studies by Tenopir and King, this study asked some general and demographic questions, and then focused on the last instance of scholarly article and book reading. The respondents answer many specific questions about the last reading, including source, time spent, purpose, and valueofthereading.thisprovidesuswitha sample of readings, in addition to a sample of readers. In spring 2011, an email invitation with an embedded link to the survey instrument was sent to all academic staff at six UK research institutions. Of the approximately 12,600 invitations distributed, 2,117 respondents answered at least one question, for an overall response rate of 16.8%. Since respondentswereallowedtoleavethesurveyat any time, skip questions, or were timed out automatically if they began the questionnaire and did not complete it, most of the questions have a lower number of responses than the total of 2,117 who answered at least one question. Academic staff members at the six UK institutions spend the largest portion of their work time on research and writing activities. Half of the respondents spend almost half (45%) of work time on research and writing. Teaching and administrative activities make up the other large sections of respondents work time, 23% and 16%, respectively. Of the 1,102 respondents who chose to give their academic discipline, about a third are either from the life or physical science fields. Humanities, social sciences, and med-

Article and book reading patterns of scholars: findings for publishers 281 ical science each account for approximately 10% and the rest of the respondents are from a variety of other academic disciplines. For analysis we collapsed the disciplines into six categories based on similarities in their fields, and redistributed the other disciplines into a corresponding category. Fine arts were combined with humanities; law, psychology, business, and education were combined with social sciences. The remaining other disciplines are interdisciplinary fields (i.e. humanities and health ) or disciplines that did not clearly fit into one of the larger categories (i.e. architecture). Three-quarters of respondents who chose to identify their age are between 30 and 60, evenly split between the three decades. The age range within each discipline has a similar distribution to the total respondents. Sixty per cent of respondents are male, which is congruent with the total UK academic staff population. 8 In past surveys of research universities and in non-university research settings, authorship was used as one measure of productivity. Over the years it has been shown that academic staff members who publish more journal articles tend to read more. 8 Over 80% of the respondents have published at least one article in a refereed journal in the past two years, and over half of staff members published at least two refereed journal articles. Approximately 70% of respondents have published at least three or more scholarly works in the past two years. Another measure of productivity is whether respondents received awards or recognition for their work. We asked if they have received any awards or recognition in the past two years; they were then prompted to describe the award or recognition. Approximately 19% of respondents report receiving awards in the past two years. Their responses range from teaching excellence recognition, grants, fellowships, best paper awards, and prizes for their research and/or work. Finding 1: articles are essential One of the questions in all of the Tenopir and King surveys from 1977 to the present is an estimate of the total number of articles read in the last month by each respondent. While it relies on personal recollection and the assumption that the last month is an accurate representation of a typical month of reading, it gives an approximation of how many articles a respondent reads, and allows us to compare over time and across populations. To aid the respondent s memory, we ask for a relatively short period of time (one month) and define articles and reading carefully. The first question states: In the past month (30 days), approximately how many scholarly articles have you read? (Articles can include those found in journal issues, Web sites, or separate copies such as preprints, reprints, and other electronic or paper copies. Reading is defined as going beyond the table of contents, title, and abstract to the body of the article). The actual number is not as important as the relative amounts among types of respondents and over time. In the past month UK academic staff read, on average, 22 articles (mean = 22.32, SD = 22.844). 13 Extrapolated to an entire year, the average number of articles read per UK academic across disciplines is 267, or 298 excluding humanities. Since the first scholarly reading surveys completed by King in 1977, we have seen a continual increase in the number of scholarly article readings for all non-humanities academics. In 1977, a survey of scientists and social scientists in the US found an average of 150 article readings per year, with an increase in each subsequent set of surveys. In 2000 03 the average reading had increased to 216 articles per year and by 2005 06 reading had increased to over 250 articles per year. 14 We looked at how the respondents spend their work time, and how their work activities may influence reading amounts. Academics who spend over half of their work time on research and writing read more articles. These research intensive academics report 26 article readings per month on average (mean = 26.09), while academics who spend less than half their time on research and writing report reading only 22 articles per month (mean = 21.40). extrapolated to an entire year, the average number of articles read per UK academic across disciplines is 267

282 Carol Tenopir, Rachel Volentine, and Donald W. King Table 1. Principal purpose of article reading by UK academic staff respondents Frequency Per cent Research and writing 862 74.3 Teaching 139 12.0 Administration 2 0.2 Current awareness/keeping up 69 5.9 Consulting, advising 10 0.9 Internal or external presentations 19 1.6 Continuing education for self 29 2.5 Engagement activities (to wider community) 4 0.3 Other 27 2.3 Total 1,161 100.0 38% of readings are very important or absolutely essential to the principal purpose On the other hand, academics who spend a majority of their work time on teaching read fewer articles. These teaching-intensive academics who spend over half their work time on teaching report 18 article readings per month on average (mean = 18.23). The medical/health science disciplines report the most article readings per month (mean = 30.64), followed by engineering/ technology (mean = 28.32), humanities (mean = 27.72), and sciences (mean = 26.60). Social scientists report the fewest article readings per month (mean = 21.39, SD=20.98). Our findings compare with past studies that show academic staff members in the medical/health disciplines read more articles but spend less time per reading compared to the other disciplines. 14 Social scientists spend on average slightly over an hour per article reading, with those in the medical/health disciplines spending 41 minutes. Scientists and humanists spend approximately 50 minutes per article reading. Academic staff members in the engineering/ technology disciplines spend, on average, the most time per reading (mean = 72.29). The survey data not only shows the importance of scholarly reading to each disciplinethroughtheusagedata(timeand amount), but also provides a picture of purpose, value, and outcomes from readings. We ask, For what principal purpose did you use, or doyouplantouse,theinformationobtained from the article you last read? Nearly three-quarters of the readings are for the principal purpose of research and writing (Table 1). The other purposes include reference, no clear principal purpose, or more than one principal purpose. 15 In relation to the principal purpose, we asked respondents to describe the value of reading by ranking the article s importance to the principal purpose and the outcome the reading has on their work. Respondents ranked the article reading on a five-point scale from absolutely essential to not at all important. Almost all of the readings are at least somewhat important (98.9%), and 38% of readings (441 of 1,160) are very important or absolutely essential to the principal purpose (Table 2). We received many comments stating the importance of scholarly article reading; many of them describing the readings as essential, critical, and could not do without them. Even when the reading is not ranked as important to principal purpose, the respondent still often mentioned that scholarly articles are indispensable or an essential source of information. The quality of the scholarly articles influences the quality of the academic work. One respondent questions, It is my understanding that scholarly articles are one of the Table 2. Importance of article reading to respondents principal purpose Frequency Per cent Absolutely essential 150 12.9 Very important 291 25.1 Important 342 29.5 Somewhat important 364 31.4 Not at all important 13 1.1 Total 1,160 100.0

Article and book reading patterns of scholars: findings for publishers 283 prime tools for conducting research and teaching. How would you do conduct research and teaching without them? Another respondent reiterates, They are absolutely 100% essential to everything I do. By understanding the essential quality of article reading, it helps establish the importance for providing access to quality articles. Finding 2: the library e-collection is the main source of articles An important part of our analysis of UK academic staff reading patterns is determining how they become aware of articles. In the survey we ask, Howdidyouorsomeoneon your behalf become aware of this last article you read? The answers reflect the many methods available to find information. We followed up the question by asking if the source they searched or browsed was a print or electronic source. For the purposes of the survey we defined browsing as without a specific objective in mind and searching as having some sort of starting point such as author s name or by subject. We included a don t know/don t remember option for staff members who may have had someone on their behalf seek out the information or who may not remember how they became aware of the article. Approximately a third of all readings reported by UK academics (32.9%) are found initially through a method of searching. Searching an electronic index or abstracting service is most common. Browsing accounted for 11% of all readings, and other, including citations, another person, or don t remember, accounted for over half (56%) of all readings. Of the articles found through browsing, 38.3% come from the library institutional subscription, most of which come from an electronic library subscription. Respondents also browsed a personal subscription (23%), an electronic search engine (14%), and a website(11%).whilewedidnotspecify whether every source is print or electronic, approximately athirdofall readings reported by UK academics (32.9%) are found initially through a method of searching Table 3. How UK academic staff obtain articles Frequency Per cent Personal subscription 56 4.7 (100.0) Print 43 (76.8) Electronic 13 (23.2) Library subscription 775 65.2 (100.0) Print 50 (6.5) Electronic 725 (93.5) Department/school 54 4.5 (100.0) Print 8 (14.8) Electronic 46 (85.2) Subject or institutional repository 26 2.2 Free web journal 109 9.2 Copy of the article from a colleague, author, etc. 66 5.6 (100.0) Print 16 (24.2) Electronic 50 (75.8) Interlibrary loan or document delivery service 25 2.1 (100.0) Print 13 (52.0) Electronic 12 (48.0) An author s website 18 1.5 Other website 32 2.7 Other source 28 2.4 (100.0) Print 4 (14.3) Electronic 24 (85.7) Total 1,189 100.0

284 Carol Tenopir, Rachel Volentine, and Donald W. King very few academics under 30 read from personal subscriptions (approximately 2%), while 10% of readings from respondents over 60 are from personal subscriptions from the data we do have, electronic sources are clearly the primary means of becoming aware of articles. Once the respondent became aware of the article, we asked where they obtained it. Almost two-thirds (65.2%) of the readings are obtained from a library subscription (Table 3). Many respondents praised the importance of library sources, including one respondent who summed it up as, Accessibility of scholarly journals and other library resources is crucial to the standing and effectiveness of a university and is a key discriminator between world-class universities and less prestigious institutions. While a personal subscription is used for approximately 5% of the readings, the readings from a personal subscription are predominantly print (76.8%), whereas 94% of the readings from a library subscription use an electronic version. On the whole, readings obtained from the library are considered more important than those obtained from other sources. Forty per cent of readings provided by the library are considered very important to absolutely essential, and less than 1% are considered notatallimportant. There are some differences between where the article is obtained and the subject discipline of the reader. While over half of the readings by each discipline are obtained from a library subscription, readings from the medical/health academic staff members report the lowest percent of readings from the library (59.3%) and the highest percent from free web journals (15.9%). Ten per cent of the readings by scientists are obtained from free web journals, while less than 6% of the article readings by the other disciplines are obtained from free web journals. While thesourcetoobtainthearticlevaries,over three-quarters of article readings in each discipline are from an electronic source. While age does not significantly influence wherethearticleisobtained,thereisan increase in the readings from personal subscriptions as the respondent s age increases. Very few academics under 30 read from personal subscriptions (approximately 2%), while 10% of readings from respondents over 60 are from personal subscriptions. At least 65% of article readings within each age group (decade) are obtained through the library, except for respondents over 60 (54.3%). The number of articles from an electronic resource exemplifies the importance of electronic journals, and shows the continual need to maintain these collections and work towards making and marketing electronic copies of scholarly journals. Finding 3: library e-collections are convenient and an easy source to find and obtain articles Academics like it when they can find a relevant article and obtain its full-text from the same source or a quick link. Library collections that enable this are valuable to academics. Regardless of how a reading is found, the majority are obtained from a library subscription. One respondent stated, A well stocked library (including e-materials) is absolutely essential to all aspects of scholarly activity. I can think of no feasible alternative. Readings found through searching or citations are more likely to be obtained through a library subscription (73.3% and 77.2%, respectively). Both becoming aware of and obtaining articles require time, and academics want to spend as little time as possible on these tasks. One respondent states, It is important to get an article almost immediately through an on-line subscription. There is far too much information out there and very little time to screen through and read articles. I find that there is less chance of reading an article, when there is a delay between the time of finding the article and getting hold of it. One of the most frequent comments we received is the importance of speed and easy access to articles. Electronic sources allow for less delay between locating and obtaining the article. Articles obtained through a print source take an average of 5 minutes to obtain, while articles from an electronic source take approximately 3 minutes to obtain. While there is a significant association between print and electronic sources of articles and the time to obtain, the individual sources (library subscription, personal

Article and book reading patterns of scholars: findings for publishers 285 subscription, etc.) do not have a significant influence on the time to obtain. All sources require 2 3 minutes to obtain the article, except for interlibrary loan, which averages 15 minutes. Clearly, academics have become accustomed to speedy access to articles once they become aware of them. Another aspect of convenience is location. While academics are using the library s resources, they are often accessing the library s resources remotely and are rarely reading in the library. UK academics do the majority of their reading in their office and laboratory (60%), and less than 2% of the readings are read in the library (Table 4). Location is no longer a major factor in a staff member s access to academic sources because the scholarly articles can be accessed and read from multiple locations. All of these features save the academic s time e-collections, synchronization of finding and obtaining services, full-text downloads, and remote access. In many ways these services are not just nice but necessary. One respondent comments, The online access provided via the university s library subscriptions to a wide range of journals across many disciplines is vital to me meeting my expected workload as a Research Associate. Another respondent reiterates, I could not do my research without the speedy and wide-ranging access to material provided through the electronic resources from the library. New bundles and services which continue to improve the functionality and efficiency of e-collections will only further benefit the academic endeavor. Table 4. Location of article reading by UK academic staff respondents Frequency Per cent Office or laboratory 699 60.2 Library 20 1.7 Home 324 27.9 Travelling 116 10.0 Elsewhere 3 0.3 Total 1,162 100.0 Finding 4: scholars consider articles important because the readings keep them up-to-date in their field, but they also want older material for reference and as a foundation for new ideas While readings principally support the scholar s research, one of the most frequent comments we received about the importance of scholarly article readings was its use in keeping their research up to date in the field. The respondents repeatedly said, They are essential for keeping up to date with new interpretations and discoveries, and to keep up to date with research in the fields I teach and conduct research. Current issues of journals are important, but older material is also crucial to scholars. One respondent points out, [It is] essential to have access to older articles to prevent reinventing the wheel. Older publications allowscholarstoseeideasortrendsevolve and develop over time, as well as to familiarize themselves with the foundation/key ideas in a field. In the surveys from 1977 to 2005 in the US, we have seen an increase in reading of articles older than the first year of publication, though reading is still skewed to the most recent articles. 16 The findings in the UK follow the same trend we first saw in the 2005 US and Australia studies. Nearly one-half of the readings are from articles in their first 18 months of publication (Table 5), yet the year of publication ranges from as early as 1890, with 13 articles published before 1950. Over two-thirds of readings are Table 5. Age of article read by UK academic staff respondents arranged by date groupings Year Frequency Per cent Over 15 years 128 11.1 (before 1996) 11 years 15 years 72 6.3 (1996 2000) 6 years 10 years 157 13.7 (2001 05) 2 years 5 years 253 22.0 (2006 09) 1 year 1½ years 539 46.9 (2010 11) Total 1,149 100.0 current issues of journals are important, but older material is also crucial to scholars

286 Carol Tenopir, Rachel Volentine, and Donald W. King Table 6. Final format of last article reading by UK academic staff Frequency Per cent Print article in a print journal 114 9.8 Photocopy or fax copy 30 2.6 Online computer screen 348 29.9 Previously downloaded/saved and read on computer screen 173 14.9 On a mobile, e-reader or tablet screen 8 0.7 Downloaded and printed on paper 476 40.9 Other 14 1.2 Total 1,163 100.0 our findings and their research further suggests that backfiles are a key investment in addition to current subscriptions academics still print-out articles for final reading, although on-screen reading is increasing within five years of the article s publication (68.9%). Humanists read a higher proportion of articles after the first year of publication, while medical/health staff members read more current articles. While the most frequent year of publication for each discipline is either 2010 (humanities) or 2011 (all other disciplines) (i.e. within the first 16 18 months of publication), half of the readings for humanities were published before 2005 (6.5 years old or older). On the other hand, half of the readings in the social sciences and engineering/technology disciplines were published after 2008 (approximately 2.5 years old), and half the readings in the sciences and medical/health disciplines were published after 2009 (no more than 18 months old). A quarter of the readings by humanists were published before 1996 (more than 15 years old), while less than 10% of readings from other disciplines are 15 years old or older. It is important to think about how the library packages are bundled because academics are reading older articles. Studies done by Guthrie, 17 Odlyzko, 18 and Herman 19 provide further research on the life of a journal article and its half-life. They found many older articles are heavily used when they are conveniently accessible; however, academics tend to cite more recent articles in order to seem current and up to date in their field. Our findings and their research further suggests that backfiles are a key investment in addition to current subscriptions. Finding 5: academics still print-out articles for final reading, although on-screen reading is increasing Although approximately 87% of readings are obtained from an electronic source, slightly less than half of all readings are read on a computer or other electronic screen (Table 6). The other half are read in a print form, Table 7. Association between age of UK academic staff respondent and reading form of article Reading form Print Printed electronic source Electronic Row total Under 30 17 81 88 186 9.2% 43.5% 47.3% 31 40 22 137 148 307 7.2 44.6% 48.2% 41 50 39 116 120 275 14.2% 42.2% 43.6% 51 60 32 81 97 210 15.3% 38.6% 46.1% Over 60 15 23 32 70 21.4% 32.9% 45.7% Column total 125 438 485 1,048

Article and book reading patterns of scholars: findings for publishers 287 either from a print journal or downloaded and printed from an electronic source (54.5%). Many respondents complained of the eye strain and formatting issues of electronic journal article reading. One respondent observes, Electronic resources have revolutionized the way research is done but journal and book formats are not designed well for reading on screen so a lot of printing resources are needed, and another respondent adds, I like reading hard copy, but now mostly read publications on computer screens, which are less comfortable on the eyes and neck. The majority of articles are from electronic sources, regardless of age of reader, but there is an increase in the use of print journals by respondents as their age increases (Table 7). Twenty-one per cent of article readings by respondents over 60 are from a print source and 33% are read from a printed copy from an electronic source. Nearly half of the respondents in each age group are reading on computer screen. While academics and libraries no longer purchase as many print journals as before, some of the costs are offset by the increased price of printing the material. One respondent says, Journal and book formats are not designed well for reading on screen so a lot of printing resources are needed. Innovations to make reading on-screen more eye-friendly may help increase the use and popularity of e-resources. Reading on-screen is increasing. In our studies in 2005 in the US only 17% of articles were read on a computer screen. 16 This change is testament to the increase in articles obtained from electronic sources, and the saturation of electronic material into all aspects of academic life regardless of age or discipline. Finding 6: books are also important for teaching and research, particularly in the humanities Readings from books also play a crucial role in academic work. One respondent comments, In my discipline both journal articles and books (monographs and edited volumes) play an important role in both teaching (which is often enquiry-based) and research, and another respondent notes, edited books and monographs remain just as important [as articles] in my field. UK academic staff average seven book or book chapter readings per month or approximately 84 readings from books per year (mean = 6.95, SD = 9.589). 20 Humanists report the most book readings by a large margin (mean = 20.50). Academics in the social sciences report an average of nine book readings per month, while those in the sciences and medical/health disciplines report the fewest readings per month (Table 8). There is a significant relationship between the respondent s age and the number of book readings (F = 1.216, P = 0.302). As the age of the respondent increases, the number of book readings increases. Respondents under 30 report approximately five book readings per month, and respondents over 60 report approximately eight book readings a month. When we compared academics who spend a majority of their work time on research and those who spend less than half their work time on research, we found a significant relationship between the amount of book readings. Research-intensive academics who spend over half their work time on research, read from fewer books or book chapters (mean = 4.12) than those who spend less than half their work time on research (mean = 8.83). On the other hand, teaching-intensive academics who spend a majority of their work time on teaching read more from books or book chapters (mean = 11.41) than those who spend less than half theirworktimeonteaching(mean=6.58). As with article readings, over half of book readings are for the principal purpose of Table 8. Average number of book readings by UK academic staff respondents discipline n No. read Social sciences 259 9.02 Sciences 366 3.04 Humanities 120 20.50 Engineering/technology 178 5.27 Medical/health 145 3.70 Others 20 7.30 Total 1,088 6.95 researchintensive academics who spend over half their work time on research, read from fewer books or book chapters

288 Carol Tenopir, Rachel Volentine, and Donald W. King Table 9. Principal purpose of book reading by UK academic staff respondents Frequency Per cent Research and writing 531 57.6 Teaching 254 27.6 Administration 1 0.1 Current awareness/keeping up 23 2.5 Consulting, advising 20 2.2 Internal or external presentations 10 1.1 Continuing education for self 48 5.2 Engagement activities (to wider community) 7 0.8 Knowledge transfer or enterprise activities 4 0.4 Other 23 2.5 Total 921 100.0 word of mouth is important for promoting books research and writing (58%, 531 of 921). Approximately a quarter of readings (28%) are for teaching and 5% are for continuing education (Table 9). The other responses include for review and more than one principal purpose. Approximately 91% of the readings (910 of 918) are considered at least somewhat important (Table 10). While academics are not reading as many books per month as articles, books still play a crucial role in their work. The books they are reading are considered, on average, extremely important and support research and teaching. Readers return to the same book and may read just sections or chapters. Scholarly books support academic work, in particular, in the humanities. Finding 7: books are often circulated between colleagues, and word of mouth is important for promoting books While articles are primarily sought out through searching, browsing, or another active means of finding, books are often Table 10. Importance of books to UK academic staff respondents principal purpose Frequency Per cent Absolutely essential 168 18.3 Very important 287 31.3 Important 274 29.8 Somewhat important 181 19.7 Notatallimportant 8 0.9 Total 918 100.0 foundbywordofmouth.overaquarterof the book readings are found through another person. In addition, a large group of readers already knew of the book; consider it a core text in the field; or have it on their bookshelf. Over a third of the book readings were found through word of mouth, including readings found by advertisements, for review, and from another person, and 8% of the readings are from books already owned. While articles seem to have a high turnover (academics are constantly reading new articles), books tend to be circulated between scholars and re-read by individuals. Many of the respondents consider the books they read a key element in personal libraryovermanyyears,a keyreference volume, or a major text book in the field. We assume, therefore, many of the readings are re-readings. Even though we did not ask the respondents for the year of publication or if they knew the information in the book prior to reading, we expect the average year of publication would be older and more respondents would know all or parts of the book prior to reading than with article readings. We asked, After you became aware of this book, from where did you obtain it? The wording was kept similar to the other sections for comparison, but the answer choices were modified to reflect the different sources for books. Thirty-nine per cent of book readings are from books that are purchased (364 of the 931) (Table 11). The library or archives collection and interlibrary loan account for a little over a quarter of book readings

Article and book reading patterns of scholars: findings for publishers 289 Table 11. How UK academic staff respondents obtain books Frequency Per cent I bought it for myself 364 39.1 The library or archives collection 239 25.7 Interlibrary loan or document delivery service 18 1.9 School or department collection 24 2.6 A colleague, author or other person provided it to me 98 10.5 A free, advance, or purchased copy from the publisher 130 14.0 Other source (58) (6.2) Online, e-book 19 2.0 Other 39 4.2 Total 931 100.0 (27.6%), and another quarter of book readings are from books provided by another person or the publisher (24.5%). The other responsesvary,but19readingswereidentified as electronic books (2%, 931). Unlike with article readings, the library s collections are not the main source of book readings. Academics may prefer purchasing books because they plan on re-reading books over time and passing them between their colleagues. It may also be that the library does not have the desired books or that print books in a personal collection are just more convenient that having to go the library for print books. Readings for research or teaching are more likely to come from the library, while readings for engagement activities, presentations, and enterprise activities come from another person (colleague or publisher). This suggests that book readings for teaching or research, the tasks that take up a larger percentage of work time, are more likely to be explicitly sought, while readings for purposes that take up less work time come from readily available sources that do not require a lot of work to find or obtain. Finding 8: e-books for scholarly reading are not yet common among academic staff Only a small percentage of scholarly book readings obtained from the library by academic staff in the UK are currently from e-books. We expect these numbers to be greater for students, but did not survey students in this study. A 2009 study in the UK found that 65% of staff and students have read an e-book for work, study, or leisure, and over half of those readings were obtained through the library (51.9%). 12 Leisure reading was specifically excluded in our study. In the open comments section of our survey some respondents praise e-books with comments such as: e-books have revolutionized the way research is done. It is much more efficient than browsing hardcopy, but journal and book formats are not designed well for reading on screen, and [e-books] would make rare and specialist books much more available. The library provision of books, in particular e-books, is clearly an area for further research. Finding 9: academics who publish more and win awards also read more In past studies, we have found that academics who have won awards or received special recognition in the past two years read more articles. 8 This is also true in the UK, as respondents who earned an award within the last two years read more articles than those who have not received an award. Academics who received an award in the past two years read on average 29 articles a month, six more than those who have not received an award. Academic staff who received an award within the past two years also report more readings from books. On average, awardwinning staff read from eight books (mean = 8.08), while non-award winning staff read from six books (mean = 6.46). The number of publications also has a significant association with the number of unlike with article readings, the library s collections are not the main source of book readings academics who publish more and win awards also read more

290 Carol Tenopir, Rachel Volentine, and Donald W. King nearly half of the respondents read, view, or participate in one or more of the social media tools article readings. Academics who publish more, including refereed and non-refereed articles, books, and conference proceedings, read more articles. Academics who report more than 20 publications in the last two years read on average 36 articles per month, while those who produced only one or two publications in the last two years read just 20 articles on average per month. Finding 10: social media supports traditional scholarly material Social media tools have developed over the past decade, and their use has undoubtedly increased in both the academic and nonacademic world. In this study we wanted to measure both amount of social media use and the number of social media tools used, and see if use of social media has an influence on reading of traditional materials. Many UK academics use one or more forms of social media for work-related purposes, but creation is less common. Use, however, is more often occasional rather than on a regular basis. Open-ended comments confirmed that social media may help spread some ideas and provoke thoughts, but are not as valuable as traditional scholarly material. One respondent says, magazines and some social media provide interesting grounds for idea but cannot at any point replace high quality peer reviewed journal articles, and another staff member believes, they are great for popularizing highlights of science but should not be allowed to skew investment away from core, peer reviewed publications. Nearly half of the respondents read, view, or participate in one or more of the social media tools at least occasionally. A variety of specific tools are used, with blogs, video, and user comments in online articles most commonly used. Participation is more frequent than creation, with only a quarter of the respondents reporting they create one or more social media tools. There is a significant association between the number of article, book, and other publication readings and the participation in most of the social media tools. Contrary to our expectations, respondents who read more articles participate in more social media tools and create more tools. Respondents who participate and create more social media tools also read more books. There are some differences in use or creation based on demographic factors. Social scientists and humanists are more likely to be bloggers. Age makes some difference, particularly for creation of micro-blogs (twitter). Videos are most often viewed or created for teaching, rather than research. High frequency users or creators of social media are more likely to be age 50 or younger and, contrary to our expectations, read more scholarly materials than others. Academics who are engaged with traditional materials for their scholarly work are also embracing various forms of social media to a higher degree than their colleagues. For now, social media is not replacing traditional scholarly resources for research and teaching, but rather enhancing their use. Conclusions Academics devote a large amount of their time to reading articles and books. These readings remain important to academics of all ages and disciplines for many purposes, notably for research and teaching. While we cannot claim cause and effect, the profile of a successful academic in the UK is one who reads more on average. Award-winning and prolific academics in the UK read more articles than their colleagues, another indirect measure of the importance of scholarly materials to the academic enterprise. There are some variations in reading patterns, depending on the personal characteristics of the reader or purpose of reading. Forexample,thenumberofarticlesorbooks read monthly, the time spent per reading, and how articles and books are discovered or obtained varies with characteristics of readers such as age and subject discipline. Readings for the purpose of research take more time and are more often come from a library s e-journal collection. Readings for current awareness are more likely to come from a personal subscription than are other readings. Reading patterns for articles and books are quite different. Academics rely on their library e-journal collections for much of

Article and book reading patterns of scholars: findings for publishers 291 their article reading, but still prefer printed books in their personal collections. E-books have not yet made much of an inroad for scholarly reading by UK academics. There is an opportunity here for both publishers and libraries. Just as e-journal collections have changed reading patterns over the last decades, allowing academics to locate and obtain more article readings more quickly, 1 convenient access to scholarly books through the university library may allow more academics to read from more books for teaching or research. The value of scholarly reading to the work of academics is high and has remained so over many decades. Understanding the different patterns of reading between scholarly articles and books, by principal purpose of reading, and by individual readers in different disciplines or of ages helps publishers and libraries together continue to provide the best access systems to the resources that are so important to academic work. References 1. Tenopir,C.,King,D.W.,Spencer,J.,andWu,Lei. 2009. Variations in article seeking and reading patterns of academics: What makes a difference? Library & Information Science Research, 31: 139 148. doi:10.1016/j.lisr.2009.02.002 2. Tenopir, C. and King, D.W. Towards Electronic Journals: Realities for Scientist, Librarians and Publishers. Washington DC, Special Libraries Association, 2000. 3. Tenopir, C., Wilson, S., Vakkari, P., Talja,S., and King, D. W. 2010. Cross country comparison of scholarly e-reading patterns in Australia, Finland and the United States. Australian Academic & Research Libraries (AARL), 41: 26 41. 4. King, D.W. and Tenopir, C. Using and reading scholarly literature. In Annual Review of Information Science and Technology 34. Medford,NJ,InformationToday, Inc., 2001, pp. 423 477. 5. Healy, L. W., Dagar, L., and Wilkie, K. M. Custom report prepared for the Digital Library Federation/Council on Library and Information Resources. Burlingame,CA, Outsell, 2002. 6. Talja, S., and Maula, H. 2003. Reasons for the use and non-use of electronic journals and databases: a domain analytic study in four scholarly disciplines. Journal of Documentation, 59, 673-691. http://www.uta.fi/~lisaka/taljamaula.pdf. 7. Vakkari, P. 2006. Trends in the use of digital libraries by scientists in 2000 2005: A case study of FinELib. In A. Grove (ed.), Proceedings 69th annual meeting of the American society for information science and technology (ASIST) 43 Austin, TX. Medford, NJ, Information Today, Inc. Retrieved 17 March 2009, from http://eprints.rclis.org/archive/00008278/ 8. King, D.W., Tenopir, C., Montgomery, C. H., and Aerni, S. E. 2003. Patterns of journal use by faculty at three diverse universities. D-Lib Magazine, 9. http://www.dlib.org/dlib/october03/king/10king.html. 9. Brown, C.M. 2003. The role of electronic preprints in chemical communication: analysis of citation, usage and acceptance in the journal literature. Journal of the American Society of Information Science and Technology, 54: 362 371. http://www.asist.org/publications/ JASIS/Best_Jasist/2004Brown.pdf. 10. Research Information Network. E-journals: their use, value and impact final report. A RIN and RLUK Report. January 2011. http://www.rin.ac.uk/our-work/communicating-and-disseminating-research/e-journals-their-use-value-and-impact. 11. Nicholas, D., Williams, P., Rowlans, I., and Jamali, H. R. 2010. Researchers e-journal use and information seeking behaviour. Journal of Information Science, 36: 494 516. doi: 10.1177/0165551510371883. 12. CIBER. JISC National E-Books Observatory Project: Key Findings and Recommendations Final Report. CIBER, 2009. Available at http://observatory. jiscebooks.org/reports/jisc-national-e-books-observatory-project-key-findings-and-recommendations/. 13. All calculations exclude outliers, unless otherwise noted. 14. Tenopir,C.,King,D.W.,Edwards,S.,andWu,L. 2009. Electronic journals and changes in scholarly article seeking and reading patterns. Aslib Proceedings: New Information Perspectives, 61: 5. doi: 10.1108/00012530910932267. 15. Tenopir, C., Volentine, R., and King, D.W. 2012. Scholarly reading and the value of academic library collections: results of a study in six UK universities. Insights: The UKSG Journal, 25July. 16. King, D.W., Tenopir C., Choemprayong S., and Wu L. 2009. Scholarly Journal information seeking and reading patterns of faculty at five US universities. Learned Publishing, 22: 2. 17. Guthrie, K. 2000. Revitalizing older published literature: preliminary lessons from the use of JSTOR. Economics and Usage of Digital Library Collections Conference. MIT Press. http://www.si.umich.edu/ PEAK-2000/guthrie.pdf. 18. Odlyzko, A.M. 2000. The rapid evolution of scholarly communication. Conference on the economics and usage of digital library collections. http://citeseerx.ist.psu.edu/ viewdoc/download?doi=10.1.1.3.8895&rep= rep1&type=pdf. 19. Herman, E. 2004. Research in progress: some preliminary and key insights into the information needs of the contemporary academic researcher. Part 2. Aslib Proceedings, 56: 118 131. doi: 10.1108/00012530410529495. 20. Mean excludes one outlier at 350. With the outlier the mean is 8.13 (SD = 16.54). Carol TENOPIR (ctenopir@utk.edu) Rachel VOLENTINE (rvolenti@utk.edu) Donald W. KING (donaldwking@gmail.com) University of Tennessee Center for Information and Communications Studies Knoxville, TN 37996, USA e-books have not yet made much of an inroad for scholarly reading by UK academics