Can in-house use data of print collections shed new light on library practices? Statistical evidence from a five-year longitudinal study in China

Abstract

Libraries worldwide are encountering limited budgets, rising e-collection prices and declining check-out use of print books. It is necessary for administrators to demonstrate physical library value to stakeholders and make empirically based collection development decisions. This five-year longitudinal study of the Southern Medical University Library in China was conducted based on the in-house use and the check-out use data of the whole collection from 2012 to 2016. We found that the in-house use of print materials was positively correlated with the check-out use. In-house use represented 26.78% of the total circulation, on average. Academic bound periodicals and reference books were rarely used in-house. The in-house use ratio differed greatly among subjects, and subjects with frequent check-out use approximated the average ratio. Collections with high check-out use also had a high in-house use ratio. Of the books, 46.1% of the titles were used only in-house, and titles of in-house use were 1.5 times that of check-out use, although the average frequency of in-house use was lower than check-out use. Check-out use data could not comprehensively represent in-house use. In-house use data provided important information for the physical library’s value estimation and collection weeding. We provide practical suggestions on in-house use data collection, collection development and weeding work.

Keywords

Check-out use circulation collection development in-house use library value weeding

Introduction

Impacted by Web search engines and information media in the rapidly developing digital information society, academic libraries are encountering unexpected challenges and competition. Brian T. Sullivan’s (2011) publication “Academic Library Autopsy Report 2050” in the Chronicle of Higher Education triggered debate about the demise of academic libraries (Li, 2013; Wang, 2012). Sullivan explained the first important reason why the academic library is dying is because book collections are becoming obsolete. Fully digitized collections of nearly every book rendered physical book collections unnecessary. Some libraries opted for subscriptions to digital books, and students could view these online via computers or portable devices at any time and in any place (Sullivan, 2011). The declining print book circulation worldwide witnessed the plight of academic libraries in recent decades. Many academic libraries bought more new and popular books to attract readers in China; furthermore, they doubted whether it was worthwhile to maintain large numbers of print collections, as physical space was limited. The evidence of the declining use of print collections was mainly based on check-out statistics from the Integrated Library System (ILS), without considering the in-house use (Rose-Wiles and Irwin, 2016). Although many libraries had noticed the in-house reading phenomenon, statistics for the in-house use of materials were not widely available and may not have been widely collected (Martell, 2008). Without considering the in-house use data, the value of a physical library may be underestimated and imprecise (Adams and Noel, 2008). Ultimately, physical book circulation remains an important aspect of a library’s value and may be positively correlated with academic success (Cetin and Howard, 2016). The declining check-out figure is alarming to library administrators who fear that support for library buildings and print collections would be reduced (Allison, 2015). Particularly in the context of declining economic growth, Chinese academic libraries are confronting reduced budgets and rising prices of digital resources. The declining check-out statistics are obviously difficult to use to persuade stakeholders to increase or maintain the library budgets allocation without considering in-house use. Statistics including check-out and in-house use can better demonstrate the underestimated value of a physical library.

In addition, the neglect of in-house use statistics also affects weeding work, which is often based on check-out use data. We are not sure whether books used in-house are simply those that are checked out. In general, collections such as periodicals, magazines and reference books cannot be taken out of libraries such that there are few usage statistics about these collections. Library administrators often consider removing bound periodicals to release more physical space. Whether libraries should cut down on print periodicals and buy electronic equivalents needs to be answered via longitudinal usage statistics. Statistics of in-house use of print materials would be very helpful for weeding books. However, little research has been conducted to provide enough statistical evidence on this issue. Thus, we adopted five-year statistics from the Southern Medical University Library in China to provide such evidence.

This study focuses on in-house use of print collections (including books, periodicals and magazines), and we also compare them with e-book use in the selected Chinese academic library. Our objective is to obtain a more comprehensive understanding of print collection usage and inform decisions about collection development, space reclamation as well as book budget allocations (Rose-Wiles and Irwin, 2016). Our study is designed to address the following questions:

RQ1. What was the proportion of in-house use of print books in Chinese academic libraries based on the whole collection over a relatively long period of observation? Were there any correlations between in-house use and check-out use?

RQ2. Were bound periodicals and reference books that were not allowed to be lent out widely used in-house?

RQ3. Did in-house use ratios differ among subjects? What kinds of subjects’ in-house use ratios could represent the whole collection?

RQ4. Were the books used in-house only those that were checked out? Did usage frequencies differ between in-house use and check-out use?

This paper is organized as follows. First, we review the relevant literature on in-house use of print materials. Next, we describe the research context and introduce our data collection and analytical techniques. We then report the results of this study, together with a discussion of findings, practical application and limitations. Finally, we conclude our research and note directions for future research.

Literature review

The in-house use of print materials could make the value of the physical library visible; as such, in-house reading has garnered librarians’ focus. Several studies focused on the in-house use of books and journals have been implemented in recent decades.

Most relevant studies mainly focused on the in-house use of print books. McGrath (1971) collected data, which represented approximately 7% of one-year use at the University of Southwestern Louisiana Library over four weeks and found that the check-out use figure could be used to predict total library usage by employing the observed ratio of out-to-in. Over-the-counter circulation represented real demand, and the subjects of books checked out were simply those requiring attention regarding collection development. Harris (1977) examined 10% samples of the general open-access bookstack from January 1973 to July 1974, and the comparison of check-out with in-house use data suggested that the in-house use numbers appeared to be very closely associated with the check-out numbers, which could be reliably counted. The proportion of the check-out use was far greater than that of the in-house use of books, which represented 40% of the total use. Harris also suggested that the check-out use figure could be used as an indicator. Kent (1979) also suggested that check-out use was a suitable measure of total use. Hindle and Buckland (1979) found that books checked out minimally garnered relatively minimal in-house use. Lane (1987) attempted to examine the relationship between the check-out and in-house use of books during 1984–1985 and found that check-out use could be a reliable indicator of books used in-house. Seith et al. (1992) presented a study of both check-out and in-house use of 13,029 volumes (randomly chosen from a collection of 1.1 million volumes) of serials and monographs in all subjects at University of California over a 7-year period. The researchers found that 11.2% of the monographs and 13% of the serial volumes were not checked out but had some in-house use. More than 30% of the monographs and 25% of the serial volumes had one kind of use but not the other. Monographs received much more check-out use, while serials received more in-house use. Weeding based on check-out use alone would eliminate at least 112,000 volumes that had actually been used very recently. This finding challenged the previous notion by Lane (1987) and McGrath (1971) that in-house use could be completely inferred from check-out use figures. Ridley and Weber (2000) simultaneously viewed students’ in-house use behavior of 115 monographs and 61 serials in a 13-week trial. The results revealed a difference between disciplines; the in-house use rates were low, ranging from 2% to 13% for different disciplines. Martell (2008) cited sporadic data from the Association of Research Libraries (ARL) supplementary statistical reports and analyzed the decline of the physical use of academic library collections between 1995 and 2006. While check-out use among ARL declined substantially during that period, the median in-house use also experienced a 57% decrease, and certain libraries were worse. Rose-Wiles and Irwin (2016) conducted a one-year study on the in-house use of certain print books and indicated that Seton Hall University patrons were using books within the library to a very significant extent despite the declining check-out use. The researchers found that nearly 30% of circulation transactions were books used in-house. Of the medical books, 46% used in-house were checked out, and 19% of science books used in-house were checked out.

Some studies focused on the in-house use of periodicals. Tibbetts (1974) gathered title and date information of periodicals used in-house during two one-week periods in April 1972 and found that a few titles represented most of the total use, and periodicals of the last five years represented 58% of the total use. Griscom (1983) collected citation data based on bibliographies in theses and dissertations at Indiana University Music Library and found that researchers in music relied on a relatively small group of periodicals. Walter (1996) gathered check-out use, in-house use, and interlibrary loan (ILL) use data of 2552 print periodicals over one year. The results showed that in-house use was higher than check-out use by ratios ranging from 4:1 to 16:1 across all subjects and publication frequencies.

Most studies noted above were published in 1970–1980, and authors only chose a small proportion of the whole collection or used a short period of time. These previous studies were mainly conducted in American or British libraries rather than developing country libraries in an electronic environment. Additionally, the studies were conducted with small sample size data that may be simplistic and inaccurate. In fact, an experiment that expects to elucidate this research question usually must meet the following two requirements. The first requirement is to begin with books that have not circulated. The second requirement is to ensure books are used in-house or checked-out over the same or similar time period. However, this assertion is impossible for most libraries since only a few libraries have instituted a system for recording in-house use. Furthermore, fewer studies focused on exploring the variability issue of the in-house use of print books in libraries over time. Therefore, we attempted to use a large sample size data of the Southern Medical University Library in China to demonstrate the value of in-house use statistics for library services.

Methodology

We conducted this study at the Southern Medical University (SMU), which is a public university located in Guangzhou city, Guangdong Province, China. SMU has two campuses, Guangzhou main campus and the Shunde off-campus. SMU is one of the top medical universities in China and is classified as a Research and Teaching Type University. SMU comprises 19 schools and many disciplines such as medicine, traditional Chinese medicine, pharmacology, biochemistry, public health, medical engineering and management. As of fall 2016, SMU had an enrollment of nearly 20,000 including more than 15,000 undergraduates and 4300 postgraduates. Freshman and sophomore students study in the Shunde off-campus, and other students study in the main campus. The collections of SMU library (SMUL) are housed in the Guangzhou campus library and Shunde campus library and comprise over 2.284 million items in total. Nearly 50% of these collections are print books, bound periodicals and SMU dissertations. There are 1.05 million e-books (mainly Chinese) and 38,000 e-journals as of fall 2016. For the past five years SMUL has invested heavily in e-resources (mainly foreign e-journals), which represented 85% of the total budgets. SMUL began to adopt a new ILS named Huiwen in November 2011. Thereafter, the Head of Circulation Service decided to collect in-house use data of print collections to evaluate the total use of the physical library. The decision was made as circulation had substantially decreased in recent years, and less than 20% of print holdings circulated every year. Huiwen provided us the opportunity to collect in-house use data of all books from 2012 to 2016.

Our study was implemented as follows. First, the Head of the Information Technology Department created an account in Huiwen and named it “in-house reading”. Then, we focused on the analysis of books and bound periodicals. SMUL subscribed to no more than 100 kinds of current journals without barcodes, for which was difficult for us to record in-house use data. We defined in-house use books as every book that had been left lying on a table, ledge, shelf, carts, sofa, chair, group study room, hidden away from their proper place and beside a photocopy machine. We pasted the instructions “Do not re-shelve the books” on bookshelves to guide students to put browsed books on carts that were arranged as one among four shelves. Despite such tips, it remained inevitable that a few students would re-shelve browsed books; such usage was not included in our study. This finding indicated that collected in-house use data would be underestimated to an unknown degree.

Next, we trained the circulation department librarians and student assistants. The training course included the basic process of collecting in-house use materials and how to scan data into the system correctly. In-house use data was collected when SUML was open except for certain holidays and vacation times, such as National day, Spring Festival, and days when the library was closed during summer/winter vacations. Librarians persistently guided students to collect books from in-house use on various floors and return them to the circulation station 10 minutes before the library’s closing time. Librarians arrived at the library 10 minutes before it was open the next morning to check again and to ensure that no books were missed. The librarians collected and scanned these books one by one into the “in-house reading” system. Thus, until the end of 2016, we had recorded more than five years’ circulation and in-house use data. In this study, we chose five-year data from 1 January 2012 to 31 December 2016 for the longitudinal study.

Finally, we processed the collected data and conducted statistical analysis. The raw in-house and check-out use data between 1 January 2012 and 31 December 2016 were downloaded by year into Microsoft Office Excel. The tabular data includes subject category, title, ISBN/ISSN, author and usage frequency. We used Excel’s pivot table function, SPSS 20 and Python programming to analyze the subjects and usage frequency. The statistical function of the Huiwen system was also used to collect all kinds of data such as subjects and patrons. To compare e-book use with print book use, we obtained the usage statistics from Superstar database’s management platform. Superstar is the largest Chinese e-book database subscribed to by SMUL, and SMUL has subscribed to approximately 580,000 books for more than 10 years.

Results

In-house use of print books and relationship between check-out and in-house use

Table 1 shows annual statistic use data of SMUL. The number of valid readers means how many readers were granted library cards; patrons number means how many readers went to the library. Total held means the number of books SMUL held; title held means the types of books SMUL held. Total in-house means how many books were used in-house; titles in-house means how many kinds of books were used in-house. Total check-out means how many books were checked out. The number of valid readers, patron number, total held and titles held by SMUL had grown 23.95%, 32.19%, 10.81% and 26.43%, respectively from 2012 to 2016. Total in-house use and titles in-house use of books had slight growth from 2012 to 2014 and then decreased sharply to lower than that of 2012. Total check-out use of books decreased each year. Table 1 tells us that the usage of print books was continuously decreasing, although valid readers and patrons were increasing.

Table 1.

Readers and in-house and check-out use of print books.

Year	Valid reader	Patronsnumber	Totalheld	Titlesheld	Totalin-house	Titlesin-house	Total check-out
2012	23,030	1,205,045	896,333	191,700	72,353	31,707	207,980
2013	26,175	1,352,839	927,237	208,096	73,866	33,976	196,934
2014	27,237	1,373,283	951,941	220,389	77,029	34,468	188,855
2015	28,003	1,334,253	975,686	232,269	66,261	31,516	185,125
2016	28,634	1,593,006	993,193	242,372	57,961	30,673	154,538

In-house use of books in this table does not include bound periodicals and reference books; for the purpose of parallel comparison, as they are not allowed to be lent out in SMUL.

In-house use of books represented approximately 25.81%, 27.28%, 28.97%, 26.36% and 25.48%, respectively, of the total circulation (check-out and in-house use) from 2012 to 2016. The proportion first increased and then decreased, with an average proportion of 26.78%. The proportion changed minimally in the past five years and was lower than the findings of Harris (1977) and Hindle and Buckland (1979) who reported the proportion approximate to be 40%. There were marginally significant positive correlations (0.05<p<0.1, p=0.085, r=0.825) between the numbers of in-house use and check-out use (see Table 2). Titles for in-house use represented only 16.54%, 16.33%, 15.64%, 13.57% and 12.66% per year, respectively, of the total titles held by SMUL. The proportion of titles used in-house also decreased gradually from 2012 to 2016.

Table 2.

Correlations between in-house use and check-out use.

Correlations
		In-house	Check-out
In-house	Pearson Correlation	1	.825
	Sig. (2-tailed)		.085
	N	5	5
Check-out	Pearson Correlation	.825	1
	Sig. (2-tailed)	.085
	N	5	5

Correlation is significant at the 0.05 level (2-tailed).

Previous studies reported e-book use was an important reason for the decrease of print book use (Bailey, 2006; Chen-Gaffey and Getsay, 2015; Martell, 2008). We compared the usage statistics of print materials with the largest Chinese e-book database, Superstar (see Table 3). Superstar homepage hits grew gradually from 13,189 to 18,274 during the period between 2012 and 2015 but decreased sharply to 12,940 in 2016. Overall, the Chinese e-book visits were growing, although the visiting number was very low compared with the valid reader number each year. There was no significant correlation between total print use and Superstar e-book use (p=0.645).

Table 3.

Usage statistic from the Superstar e-book database.

Year	Homepage hits
2012	13,189
2013	18,662
2014	14,821
2015	18,274
2016	12,940

In-house use of bound periodicals and reference books

Bound periodicals and reference books were not allowed to be taken out; therefore, we obtained no usage statistics from the ILS. These collections occupy much physical space in most Chinese academic libraries. It was necessary to collect in-house use data to make scientific weeding decision. SMUL held 44,158 English bound periodicals, 94,282 Chinese bound periodicals and 9756 reference books in 2016. As SMUL only subscribed few periodicals and reference books since 2007, we ignored the increased amount every year. The average usage number of English bound periodicals was 25, while that of Chinese bound periodicals and reference books were 252 and 184, respectively, per year in the past five years (see Table 4); this meant the in-house use proportions for bound periodicals and reference books (also the total use proportion) were only 0.06%, 0.3% and 1.9%, respectively. The proportions were far lower than that of print books used in-house. The usage of English bound periodicals was exceptionally low, and most Chinese bound periodicals were used for leisure.

Table 4.

In-house use of bound periodicals and reference books.

Year	Englishperiodicals	Chinese periodicals	Reference books
2012	16	146	182
2013	37	143	194
2014	29	497	246
2015	32	297	205
2016	11	179	95
Average	25	252	184

Subject distribution of books used in-house

We examined the in-house and check-out use of different subjects using the Chinese Library Classification (CLC) (see Table 5). There were six frequently used subjects, Medicine (R), Literature (I), Language (H), Philosophy & Religion (B), History & Geography (K), and Economics(F), both for in-house and check-out use. These subjects represented 75%–80% of circulation both in and out of the library (see Appendix A). The in-house use proportions differ much among subjects in different years (see Appendix B). In-house use proportions of the six subjects noted above were most approximate to the average proportion (see Appendix B).

Table 5.

Subject distribution of books used in-house and checked-out.

CLCletter	2012		2013		2014		2015		2016
CLCletter	I^*	C^*	I^*	C^*	I^*	C^*	I^*	C^*	I^*	C^*
A	175	588	219	998	244	882	178	914	177	979
B	4656	12,797	4282	10,640	4333	10,970	4017	10,840	3223	9491
C	1476	3734	1446	3862	1409	3454	1305	3901	1049	3391
D	2170	6284	2047	5203	2135	5177	1539	5091	1458	4228
E	182	347	215	239	123	174	90	185	71	139
F	3183	7439	3038	6475	2673	5076	2007	4845	1498	3500
G	1350	3328	1341	2823	1202	2488	1010	2602	1001	2704
H	7857	26,001	7338	22,571	6294	18,007	5038	16,905	4744	14,737
I	16,157	44,400	16,871	46,477	18,715	50,519	17,049	52,290	14,311	46,399
J	1849	3523	2101	4157	3226	5394	3134	4873	2369	4155
K	4535	9919	4464	8254	4594	8559	3366	7557	2731	6295
N	247	429	207	314	285	240	200	231	123	258
O	1278	3224	1425	3403	1327	3022	1419	3396	1289	3208
P	69	83	103	114	93	98	100	140	127	168
Q	894	4123	1055	4245	1470	4423	1391	4561	1217	4619
R	23,055	71,567	24,769	67,760	25,897	62,825	21,595	59,446	19,807	57,767
S	79	101	64	66	61	66	73	64	92	128
T	2663	9426	2464	8809	2519	7006	2438	6789	2371	6852
U	16	14	5	8	20	13	10	7	13	17
V	4	9	8	18	9	13	7	18	13	22
X	62	71	52	45	27	28	29	34	17	28
Z	392	573	346	453	371	421	265	436	260	404
Total	72,353	207,980	73,866	196,934	77,029	188,855	66,261	185,125	57,961	169,493

I* represents in-house use, C* represents check-out use; CLC letters: A=Marxism, Leninism; B=Philosophy & Religion; C=General Social Sciences; D=Political Science & Law; E=Military Science; F=Economics; G=Cultural, Education, Sports; H=Language; I=Literature; J=Arts; K=History & Geography; N=General Science; O=Mathematics, Physics & Chemistry; P=Astronomy & Earth Sciences; Q=Biology; R=Medicine; S=Agriculture; T=Technology; U=Transportation; V=Aerospace; X=Environmental Science; and Z=General Works.

We also tested the correlation between in-house use and check-out use of different subjects. Significantly positive correlation (p<0.01, p=0.000, r=0.997) was found (see Table 6). The findings also indicate that collections with high check-out usage also had high in-house usage.

Table 6.

Correlations between in-house and check-out use of different subjects.

Correlations
		In-house	Check-out
In-house	Pearson Correlation	1	.997^**
	Sig. (2-tailed)		.000
	N	115	115
Check-out	Pearson Correlation	.997^**	1
	Sig. (2-tailed)	.000
	N	115	115

Correlation is significant at the 0.01 level (2-tailed).

Usage frequency of books used in-house and checked-out

Until December 2016, SMUL held 242,372 kinds of print books. In the past five years, 76,438 titles were used in-house, representing 31.6% of the total titles. A total of 48,671 titles were checked out, representing 20.1% of the total titles. Combining in-house and check-out use together, approximately 90,297 titles were used during 2012–2016, representing 37.3% of the total titles. Among the total circulated books, 34,812 titles were used both in-house and checked out, which meant that a large proportion of books with in-house use (41,626 titles) were not checked out, and many titles with check-out use (13,859 titles) were not used in-house during 2012–2016. Check-out use data could not fully reflect the value of the physical library. Although books for check-out use were 3 times as numerous as those for in-house use, the titles of books for in-house use were nearly 1.5 times as numerous as those for check-out use.

Among the 76,438 titles used in-house, approximately 74.79% were used less than five times, which meant they were used no more than once per year. Of the titles, 92.1% were used no more than twice per year. Only 7.9% of the titles were used more than twice per year, on average (see Table 7). The top two books, Human Anatomy Atlas and Systematic Anatomy, were used 1020 and 924 times, respectively.

Table 7.

Usage frequency of books used in-house and checked out.

Frequency	I*	% of I*	C*	% of C*
1	28,522	37.31	8469	17.40
2	14,381	18.81	7004	14.39
3	8642	11.31	3461	7.11
4	5621	7.35	3481	7.15
5	3949	5.17	1969	4.05
6	2928	3.83	2432	5.00
7	2148	2.81	1773	3.64
8	1789	2.34	1607	3.30
9	1351	1.77	1382	2.84
10	1072	1.40	1626	3.34
11–20	4279	5.60	6971	14.32
21–30	926	1.21	3358	6.90
31–40	334	0.44	1838	3.78
41–50	170	0.22	1000	2.05
51–60	88	0.12	580	1.19
61–70	61	0.08	384	0.79
71+	177	0.23	1336	2.74
Total	76,438	100	48,671	100

I^* represents in-house use; C^* represents check-out use.

Among the 48,671 titles checked out, approximately 46.05% were used less than five times, which meant that these books were used no more than once per year. In total, 68.22% titles were used between 1 and 10 times, which meant that they were used no more than twice per year, and 31.78% titles were used more than twice per year on average (see Table 7). The top eight books were used more than 900 times. Seven of these books are textbooks titled Medical Statistics (930 times), Guide for Examination of Pathophysiology (972 times), Pharmacology (1081 times), Pathology (1139 times), Pathophysiology (1309 times), Physiology (1369 times), and Human Anatomy Atlas (1628 times) and a Chinese novel titled The Tibet Code (987 times).

Statistics in Table 7 indicate the usage frequency of check-out books was higher than in-house use. The proportion of books used more than five times during 2012–2016 in-house (20%) were much lower than that of check-out use (50%). Furthermore, we assumed that the 34,812 books were very valuable for reading as they were used both in and out of library. The average usage frequencies of check-out and in-house use were 17.56 and 6.04 per book, respectively. Comparing the usage frequency of the same book used both in and out of library, we found a significant positive correlation (p<0.01, r=0.705), which meant that the more a book was checked out, the more it would be used in-house (see Table 8). The average usage frequency of the 13,859 books that were used only by check-out was 5.34, while the frequency was 2.993 for the 41,626 books used in-house. In general, comparison of the usage frequency indicated books that were checked out would have more opportunities to be used in the future.

Table 8.

Correlations between the same books used in and out of library.

Correlations
		Check-out	In-house
Check-out	Pearson Correlation	1	.705^**
	Sig. (2-tailed)		.000
	N	34,812	34,812
In-house	Pearson Correlation	.705^**	1
	Sig. (2-tailed)	.000
	N	34,812	34,812

Correlation is significant at the 0.01 level (2-tailed).

Discussion

Prior literature findings based on a small selected sample or a short period’s observation in American or British libraries acknowledged the value of the in-house use of print books. However, there are few studies based on a large and whole collection sample, an Asian country context, strictly controlled principles of data collection and data selection, longitudinal observation and detailed comparison of in-house use with check-out use. We conducted our research by selecting five years of in-house and check-out use data during the same period from the whole collection in SMUL. This section presents our findings gained from data analysis, as follows.

In-house use and its relationship with check-out use

Statistics from the five-year study indicated that both in-house and check-out use of print books had a decreasing trend. In-house use of print books occupied a relatively stable proportion of total usage, and it had a positive correlation with check-out use. We could estimate the total circulation from check-out use data. The average in-house use proportion (26.78%) in our study was obviously lower than data from several studies in America (Harris, 1977; Rose-Wiles and Irwin, 2016). Although library administrators did their best to create a comfortable environment and developed many reading promotion activities or services such as a cafe, art gallery, computer labs, exhibitions, book recommendations, unfortunately these measures only increased patrons but could not change the decreasing trend of print book usage. In this context, in-house use statistics become increasingly more important to demonstrate the value of physical libraries. Due to the environment construction in SMUL, patrons were increasing gradually while print book usage was undergoing a continual decline during 2012–2016. A comfortable learning space is beneficial to attract more users to visit the physical library.

In contrast to the decreasing trend of in-house use and check-out use of print books, e-book homepage visits increased. However, the Chinese e-book homepage visits were not satisfying as every reader visited the homepage less than once every year, although the number grew gradually. There were no causal relationships between print book use and e-book use of Chinese books. A decrease in print book use did not necessarily lead to an in increase in e-book usage.

Academic bound periodicals and reference books were rarely used in the library

Bound periodicals and reference books were not allowed to be borrowed in many libraries, and few libraries had counted in-house use of these collections; therefore, usage of these collections was often neglected. In the digital era, whether bound periodicals merit occupying a huge physical library space is often argued in China. Data collected to support a scientific weeding decision remain inadequate. Our statistics above showed evidence that many bound periodicals and reference books should be removed to compact the stacks and release more learning space. In the past five years, only 25 English bound periodicals, 252 Chinese bound periodicals and 184 reference books were used in-house per year. This figure was surprisingly low, and most Chinese journals used in-house were magazines for leisure. Although a small number of bound periodicals and reference books may be needed one day, it remains unreasonable that SMUL used three floors to store these collections. In the future, academic libraries in China will be required to accord priority to remove academic bound periodicals and reference books when considering daily weeding work.

In-house use ratios differed greatly among subjects

Our statistics indicated that in-house use ratios differed among different subjects. Six subjects, Medicine, Literature, Language, Philosophy & Religion, History & Geography, and Economics, were the most frequently used subjects both in and out of library. These collections represented 75%–80% of the total usage. This result will provide data evidence to facilitate user-centered construction and development of library collections. Furthermore, the subjects’ distribution of books used in-house were consistent with that of check-out use, which meant that subjects with a high check-out usage ratio would also be frequently used in-house. We could estimate the subject distribution of in-house use from check-out use data.

It was not surprising that Medicine was the most frequently used subject because our research focused on a medical university library. Literature was the second most popular subject, which meant that leisure reading was also important among Chinese students. Learning English is a prevalent phenomenon, which explains why Language was the third most popular subject. The in-house use proportion of each of the six subjects was approximate to the average proportion of total in-house use (see Appendix B). We could indicate the total in-house use proportion from the subjects that were most frequently checked out.

In-house use data conveyed additional information for weeding work

As our statistics show above, the total circulation could be estimated from the check-out use figure, and the ratio of in-house use may vary in different subjects. We could choose the in-house use ratio of subjects that were most frequently checked out as a sample to predict the general in-house use ratio. The more a book was checked out, the more it was used in the library. The average usage frequency of books checked out was higher than that of in-house use. Furthermore, in-house use was not fully represented by check-out use. Only 20.1% of the total titles were checked out, while 31.6% were used in the libraries. More than 54.46% of the books used in-house had not been checked out in the five-year long period, and they were used in-house 2.9 times per book. Considering that more than 60% of the physical collections were not used at all within the five years, the 41,626 kinds of books used only in-house, which represented 46.14% of the total, were very important. The statistics above also indicated that detailed in-house use data and a relatively long period of observation should be seriously considered for weeding collections, which was often merely based on check-out use data. The statistic that only 20.1% titles were checked out in five years would greatly affect stakeholders in making decisions about allocation of library funds. In addition to the ratio of in-house use, library administrators need to provide more detailed in-house use data to demonstrate the value of the physical library and guide the practice of scientific weeding.

Practical implications, limitations and future work

Practical implications of this study

Libraries should count the in-house use of print books into the total circulation when informing the stakeholders of the usage of the physical library and print books. Considering the feasibility, other libraries could choose subjects that were borrowed most frequently as a sample to collect in-house use data for identifying the in-house use ratio of their own library. While many libraries began to reduce investment in print books and to focus on online resources as SMUL did, Allison (2015) noted that borrowing print books appeared to foster a long-term pattern of behavior. Allison noted (2015: 38) that “students who checked out materials in one year will return to check out materials in the next year, but there was less evidence that database use correlated with return sessions”. Librarians need to reconsider the collection development policy, which leads to purchasing large quantities of non-circulation books. Librarians also need to eliminate the depression regarding the decreased use of print books and physical library, focus on patron-driven acquisition (PDA), monitor the quality of book procurement, and meet readers’ needs as much as possible. Librarians are required to consider whether it is necessary to increase the number of copies of books that were used frequently.

The analysis of subject distribution indicated that print books used in and out of library were concentrated on certain core subjects. This finding is helpful for us to decide where to put a small cart to collect in-house use books and provides a reference for scientific collection development. It is necessary for librarians to exert more effort to study readers’ specific needs and user behavior in these key subjects. Furthermore, limited physical library space makes collection weeding a routine task. Weeding was generally based on check-out statistics in many libraries. No one knows whether non-borrowed books were also not used in-house. Our findings discussed above indicated that most books were used only once or twice in five years in our study. Therefore, library collection weeding should be based on a relatively long period of circulation. In general, the more a book was checked out, the more it would be used in-house. However, 46.14% titles of books continued to be used only in the library with a lower average usage frequency. Thus, combining in-house and check-out use data will make collection weeding more scientific and avoid unnecessary collection waste. Furthermore, the low usage makes it necessary to remove academic bound periodicals and reference books to compact a stack when more physical library space is needed.

Encouraging users to return to the physical library for an increase of print collections use may fail if collections do not meet users’ needs. Besides providing more popular print books, librarians are required to focus more on patron numbers and make the physical library a more convenient learning center. Librarians can explore in-house support for electronic resources and the indispensability of online access through the library for those users who cannot use e-resources remotely. Print collection use is simply one aspect of the value of a physical library, and library patrons are also important to the physical library. Our statistics indicated that the number of patrons could grow, although the print collection usage experienced continued to decline. Libraries need to take more actions to increase the attractiveness of the physical library.

Limitations and future work

This study has certain limitations in terms of the measurement of in-house use of print collections. First, books may be returned immediately to their proper place after browsing. Second, books that we collected in the libraries may be used by more than one user on one day, as they may be left on a table by one user and used later by another user. In addition, it is possible that users may take books away from shelves and then leave them on tables without actually having used them (Harris, 1977). Third, due to operational difficulties, we did not collect data about in-house use of a few current periodicals and magazines that were actually frequently used in-house, according to our daily observations. Four, we ignored the catalytic impacts of reading promotion measures when analyzing the circulation trend, such as no limits on borrowing quotas in SMUL since 2010. Finally, we did not distinguish the difference between check-out and in-house reading when considering usage of collections. The use of a book in-house must be different from that of check-out. However, this study provided valuable information on in-house use of print books based on a long period of data collection. The in-house reading behavior, the in-house use of current print periodicals and magazines, and the value of in-house use data for empirically based collection development need to be explored in the future.

Conclusion

Based on our pre-defined and strict principles for data collection, this longitudinal study focused on in-house use of print collections in a Chinese academic library. The average in-house use of print books represented approximately 26.78% of the total circulation during 2012–2016. The proportion was lower than previous American studies, which reported the proportions were 30% (Rose-Wiles and Irwin, 2016), 34% (McGrath, 1971) and 40% (Harris, 1977). Statistics also indicated that e-book use was growing; however, there was no significant correlation between print use and electronic use in the academic library. Bound periodicals and reference books were rarely used in libraries by academic users. The ratios of out-to-in differed greatly among subjects, and the ratios of most frequently checked-out subjects approximated the general ratio. Although earlier studies concluded that there was no need to collect in-house use data because of its correlation with check-out circulation, our research results revealed that in-house use was not fully reflected by check-out use. Many books without check-out circulation were used in-house. Titles of in-house use were 1.5 times that of check-out use. Our specific study on in-house use of print collections in China will be beneficial to understanding the print reading trend, conducting scientific collection development and weeding work, and providing a more accurate picture of library value.

Footnotes

Appendix

Appendix B.

In-house use proportions of different subjects.

CLC	2012			2013			2014			2015			2016
CLC	I^*	C^*	I^/(I^+C^*)	I^*	C^*	I^/(I^+C^*)	I^*	C^*	I^/(I^+C^*)	I^*	C^*	I^/(I^+C^*)	I^*	C^*	I^/(I^+C^*)
A	175	588	22.94%	219	998	18.00%	244	882	21.67%	178	914	16.30%	177	979	15.31%
B	4656	12,797	26.68%	4282	10,640	28.70%	4333	10,970	28.31%	4017	10,840	27.04%	3223	9491	25.35%
C	1476	3734	28.33%	1446	3862	27.24%	1409	3454	28.97%	1305	3901	25.07%	1049	3391	23.63%
D	2170	6284	25.67%	2047	5203	28.23%	2135	5177	29.20%	1539	5091	23.21%	1458	4228	25.64%
E	182	347	34.40%	215	239	47.36%	123	174	41.41%	90	185	32.73%	71	139	33.81%
F	3183	7439	29.97%	3038	6475	31.94%	2673	5076	34.49%	2007	4845	29.29%	1498	3500	29.97%
G	1350	3328	28.86%	1341	2823	32.20%	1202	2488	32.57%	1010	2602	27.96%	1001	2704	27.02%
H	7857	26,001	23.21%	7338	22,571	24.53%	6294	18,007	25.90%	5038	16,905	22.96%	4744	14,737	24.35%
I	16,157	44,400	26.68%	16,871	46,477	26.63%	18,715	50,519	27.03%	17,049	52,290	24.59%	14,311	46,399	23.57%
J	1849	3523	34.42%	2101	4157	33.57%	3226	5394	37.42%	3134	4873	39.14%	2369	4155	36.31%
K	4535	9919	31.38%	4464	8254	35.10%	4594	8559	34.93%	3366	7557	30.82%	2731	6295	30.26%
N	247	429	36.54%	207	314	39.73%	285	240	54.29%	200	231	46.40%	123	258	32.28%
O	1278	3224	28.39%	1425	3403	29.52%	1327	3022	30.51%	1419	3396	29.47%	1289	3208	28.66%
P	69	83	45.39%	103	114	47.47%	93	98	48.69%	100	140	41.67%	127	168	43.05%
Q	894	4123	17.82%	1055	4245	19.91%	1470	4423	24.94%	1391	4561	23.37%	1217	4619	20.85%
R	23,055	71,567	24.37%	24,769	67,760	26.77%	25,897	62,825	29.19%	21,595	59,446	26.65%	19,807	57,767	25.53%
S	79	101	43.89%	64	66	49.23%	61	66	48.03%	73	64	53.28%	92	128	41.82%
T	2663	9426	22.03%	2464	8809	21.86%	2519	7006	26.45%	2438	6789	26.42%	2371	6852	25.71%
U	16	14	53.33%	5	8	38.46%	20	13	60.61%	10	7	58.82%	13	17	43.33%
V	4	9	30.77%	8	18	30.77%	9	13	40.91%	7	18	28.00%	13	22	37.14%
X	62	71	46.62%	52	45	53.61%	27	28	49.09%	29	34	46.03%	17	28	37.78%
Z	392	573	40.62%	346	453	43.30%	371	421	46.84%	265	436	37.80%	260	404	39.16%
Total	72,353	207,980	25.81%	73,866	196,934	27.28%	77,029	188,855	29.09%	66,261	185,125	26.36%	57,961	169,493	25.48%

Acknowledgements

We thank all the librarians working in the circulation department of SMUL for their five years’ persistence in collecting in-house use data. We would like to express our thanks to Jihong Xi, Yuling Geng, Li Chen, Dan Zhao, and Jiangtao Chen for providing us different kinds of data. We also thank the editors and anonymous reviewers for their thoughtful comments and constructive suggestions that helped us to improve the manuscript.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclose receipt of the following financial support for the research, authorship, and/or publication of this article: This article is supported by Major Project of the National Social Science Foundation of China (No. 15ZDB126).

Author biographies

Xi Han is a PhD candidate in the School of Information Management, Nanjing University, China. His research focuses on library management and information behavior.

Min Song is a librarian at the Southern Medical University of China. She received her MSc in Computer Science from Wuhan University of Technology. Her research interest is in user seeking behavior.

Chunqiu Li is a PhD candidate in Graduate School of Library, Information and Media Studies, University of Tsukuba, Japan. Her research focuses on information retrieval and digital library.

Qinghua Zhu is a professor in the School of Information Management, Nanjing University, China. His research interests include information retrieval and user behavior.

References

Adams

Noel

(2008) Circulation statistics in the evaluation of collection development. Collection Building 27(2):71–73.

Allison

(2015) Measuring the academic impact of libraries. portal: libraries and the academy 15(1): 29–40.

Bailey

(2006) Electronic book usage at a Masterʼs level I university: A longitudinal study. Journal of Academic Librarianship 32(1): 52–59.

Cetin

Howard

(2016) An exploration of the relationship between undergraduate studentsʼ library book borrowing and academic achievement. Journal of Librarianship & Information Science 48(4): 382–388.

Chen-Gaffey

Getsay

(2015) More e-books, less print? What does usage data tell us? Library Collections, Acquisitions & Technical Services 39(3–4): 59–67.

Griscom

(1983) Periodical use in a university music library: A citation study of theses and dissertations submitted to the Indiana University School of Music from 1975–1980. The Serials Librarian 7(3): 35–52.

Harris

(1977) A comparison of issues and in-library use of books. Aslib Proceedings 29(3): 118–126.

Hindle

Buckland

(1979) In-library book usage in relation to circulation. Collection Management 2(4): 265–277.

Kent

(1979) Use of library materials: The University of Pittsburgh study. College & Research Libraries 40(6): 557.

10.

Lane

(1987) The relationship between loans and in-house use of books in determining a use-factor for budget allocation. Library Acquisitions Practice & Theory 11(2): 95–102.

11.

(2013) The future of academic libraries in the digital age. In: Trends, Discovery, and People in the Digital Age. Cambridge: Chandos, pp. 253–264.

12.

McGrath

(1971) Correlating the subjects of books taken out of and books used within an open-stack library. College and Research Libraries 32(4): 280–285.

13.

Martell

(2008) The absent user: Physical use of academic library collections and services continues to decline 1995–2006. Journal of Academic Librarianship 34(5): 400–407.

14.

Ridley

Weber

(2000) Toward assessing in-house use of print resources in the undergraduate academic library: An inter-institutional study. Library Collections, Acquisitions, and Technical Services 24(1): 89–103.

15.

Rose-Wiles

Irwin

(2016) An old horse revived? In-house use of print books at Seton Hall University. Journal of Academic Librarianship 42(3): 207–214.

16.

Seith

Kolle

Briscoe

(1992) The use of books within the library. Standing Orders 53(3): 197–205.

17.

Sullivan

(2011) Academic Library Autopsy Report, 2050. Chronicle of Higher Education. Available at: http://www.chronicle.com/article/Academic-Library-Autopsy/125767 (accessed 8 January 2018).

18.

Tibbetts

(1974) A method for estimating the in-house use of the periodical collection in the University of Minnesota Bio-Medical Library. Bulletin of the Medical Library Association 62(1): 37–48.

19.

Walter

(1996) A journal use study: Checkouts and in-house use. Bulletin of the Medical Library Association 84(4): 461–467.

20.

Wang

(2012) Some reflection aroused by ʻAcademic Library Autopsy Report, 2050ʼ. Information & Documentation Services 33(2): 32–34.