Commentary on the 1998-1999 CARL Statistics: An Introduction and Retrospective Overview
Dr. David Holmes
May 2000
General observations
There are certain inherent difficulties in making inter-library comparisons from annual statistics of the type collected by CARL and other library organizations. Firstly there is the philosophical problem of what to count. The things that we count are sometimes crude surrogates for the quality of a library in the context of the clientele that it serves. Unfortunately, measuring what matters: the frequency with which patrons get what they want, when they want it, at the lowest possible cost to patron and institution, is not easily measured.
Then there are the practical problems. Despite the best efforts of the compilers, it is inevitable that there are different interpretations of the definitions (what is a reference transaction, etc.?). There are different ways of counting (ab initio inventories, linear estimates, etc.) and different internal practices (loan policies, etc.). The reader should also recognize that some of the numbers reported, are estimates. For example volume counts are not usually generated from a complete inventory each year, but are usually estimated in some way.
On the other hand, some measures such as expenditure figures, can yield valid comparisons, especially when compared to overall university expenditures or student enrolments. As well, year over year comparisons of the same measure, or ratios of measures, for a given library should yield valid information about changes in that library.
In an attempt to improve the value and validity of the annual statistics and to reflect changing circumstances, CARL modified and extended its annual statistical questionnaire for the 1996-97 statistics published in May 1998. The present publication represents the third annual statistical report using these new definitions. Some data in the new statistics were not previously collected or were collected using a different set of definitions. This means that retrospective comparisons going back further than 1996-97 have to be made with care.
Finally, when making year over year comparisons of averages, it is only valid to do so for libraries that consistently reported data in the same categories for both years.
It is with the above provisos in mind that the reader should use the data presented in the annual statistical tables.
1.0 EXPENDITURES
1.1 Collections
Figure 1 shows the average total materials expenditure for the five regions (universities only) and for Ontario without Toronto. As Toronto is so much bigger and better-funded than all the other CARL libraries, the numbers for Ontario are presented with and without Toronto. In 1998-99 the highest average expenditure was in BC, closely followed by Ontario, with the lowest being in the Atlantic Provinces. Toronto leads the group with expenditures of $21.0 million, followed by the University of British Columbia (UBC), Alberta and McGill in descending order. With annual materials expenditures of $9.4 million, CISTI ranks just behind Alberta. With Toronto excluded from the Ontario average, the Ontario figure would drop to $5.2 million, dropping Ontario to third place amongst the regions.
Over the past five years the relative expenditure position of the five regions has remained the same. If the general Canadian Price Index (CPI) is used as a deflator, it would appear that all regions have more than kept up with inflation. Adjusted for general inflation, BC spent 121.3% of what it spent five years earlier on library materials (including binding and repairs). At the other end of the scale Québec spent 104.3% more.
However, a more realistic price index to use would be one that tracks the price of library materials. For Canadian research libraries there are three main factors in the changing price of library materials:
- inflation in journal prices;
- inflation in book prices;
- the declining value of the Canadian dollar relative to the United States (US) and European currencies, since the majority of CARL purchases are imported.
Clearly, a library price index should take all these into account. The best index that we have is that published by Rowe Canada (formerly Faxon Canada) who produce an annual periodical price index based on a basket of 4,444 representative periodicals from around the world in Canadian dollars. While this index does not fully represent the typical CARL collection cost, which include a variety of materials, it does track inflation in the biggest component of library collections expenditures and does take into account currency fluctuations. Because monograph inflation has tended to lag behind inflation in serials this index may slightly overstate overall library materials inflation. On the other hand, the Faxon index may under-represent the very expensive (mostly foreign) specialized research journals bought by CARL libraries. On balance it is probably a fairly good approximation of rising costs. The Faxon index shows that it cost $1.65 in 1998 to buy what $1.00 bought in 1994. Table 1 shows the dramatic effect of this inflation by region.
Table 1: Effect of Inflation on Total Materials Budget
Region |
1994-95 Average Expenditure |
1998-99 Average Expenditure |
1998-99 Expenditure Deflated by Faxon Index |
% 5-Year Drop in Purchasing Power |
BC |
$5,518,920 |
$7,129,058 |
$4,324,486 |
21.6% |
Prairies |
$4,794,440 |
$5,589,434 |
$3,390,551 |
29.3% |
Ontario |
$5,340,020 |
$6,813,414 |
$4,133,017 |
22.6% |
Québec |
$4,500,910 |
$4,966,609 |
$3,030,943 |
32.7% |
Atlantic |
$3,263,970 |
$3,688,454 |
$2,225,284 |
31.8% |
CISTI |
$8,343,950 |
$9,429,112 |
$5,719,699 |
31.4% |
Despite an increase in dollars spent on collections in all regions of the country, average effective purchasing power has dropped by between 21.6% and 32.7% by region over the last five years.
As inflation has gradually eroded purchasing power for library materials, libraries have tended to shift money out of monograph purchasing into the purchase of serials in an attempt to offset serial inflation. In 1994-95 the ratio of one-time expenditures (mostly monographs) to recurring (serial) expenditures ranged from 0.35 in Québec to 0.58 in Ontario. By 1998-99 the ratio had fallen to 0.24 in the Atlantic and 0.40 in Ontario. Only in Québec did the ratio remain constant at 0.35. The average ratio for the country as a whole has declined steadily every year from 0.46 in 1994 to 0.34 in 1998.
1.2 Overall Expenditures
In unadjusted dollars, total operating expenditures fell over the period 1994-95 to 1998-99, at 14 of the CARL universities: Calgary, Regina, Carleton, Guelph, McMaster, Ottawa, Waterloo, Western, Laval, Montreal, UQAM, Sherbrooke, Dalhousie and New Brunswick. Total library operating expenditures fell on average in the Prairies and Québec, while rising in the other three regions. CISTI’s overall operating expenditures rose while the NLC’s budget fell.
In some cases the drop in operating expenditures was quite significant. At Laval operating expenditures dropped by 23.6% over the five-year period. At UQAM it was 23.4%, at Montreal 17.6% and at Carleton 12.3% over the five-year period. On the other side of the financial coin, operating expenditures at Toronto rose by 20.4% and at Queen’s 12.0%.
As has been discussed above, if these expenditure figures are adjusted to reflect inflation in materials costs and other expenditures, they show quite a dramatic loss in purchasing power across the entire membership of CARL.
1.3 Expenditures by Enrolment
CARL periodically reports ratios where library statistics are compared to total university operating expenditures and to FTE student enrolments. This eliminates the effects of local changes in enrolment and general university finances and reflects the fact that CARL universities vary widely in size and financial strength. At the time of publication the authoritative university operating expenditure information for 1998-99 from the Canadian Association of University Business Officers (CAUBO) was unavailable. However, official student enrolment figures from Statistics Canada were available. (Note: The FTE numbers used by CARL are derived from the Statistics Canada head count data. They weight undergraduate and graduate full-time enrolments equally and pro-rate both undergraduate and graduate part-time enrolments as 1:3.5 of a FT enrolment. A more sophisticated analysis might weight graduate students more heavily and take account of other differences between categories of student.)
When total material expenditures are compared to FTE enrolments there is significant variation between institutions. UQAM had the lowest ratio in 1998 ($131.18/FTE), compared to $148.38/FTE in 1994. At the other end of the scale, Toronto claimed first place in 1998 spending $499.15/FTE, up from $350.86/FTE in 1994. The overall average for the university members of CARL was $319.02/FTE in 1998, up from $253.24/FTE in 1994.
In 1994 the five highest spenders were: Alberta, Dalhousie, Toronto, Queen’s and BC in that order. In 1998 the list was: Toronto, Queen’s, Simon Fraser, Dalhousie, BC. In 1998 the bottom three were: UQAM, Concordia and Laval in ascending order. In 1994 the bottom three were: New Brunswick, UQAM and York.
Table 2: Materials Expenditure per FTE student by Region
| Region | 1998-99 Avge. Materials Expenditure ($/FTE) | 1994-95 Avge. Materials Expenditure ($/FTE) | 1998-99 Avge Materials Expenditure deflated by Faxon Index ($/FTE) |
|---|---|---|---|
Atlantic |
$348.75 |
$282.50 |
$211.55 |
Québec |
$232.95 |
$203.84 |
$141.31 |
Ontario |
$356.34 |
$257.53 |
$216.16 |
Prairies |
$320.64 |
$263.08 |
$194.50 |
BC |
$386.51 |
$327.19 |
$234.46 |
All CARL Universities |
$319.02 |
$253.24 |
$193.52 |
Table 3 shows how the regional/provincial/national average expenditures per student compare and how they have changed over five years. Despite apparent increases in average materials expenditures per student over the past 5 years, the reality is that actual purchasing power per student has dropped throughout the country. Measured against the Faxon Canada serials price index, average materials expenditure per student has dropped 23.6% in terms of what it bought in 1994.
Total expenditures on libraries in 1998-99 ranged from $642/FTE in Québec to $1,008/FTE in BC. However when general inflation is taken into account, as measured by the CPI, actual support for libraries declined in Québec, and rose only slightly in the other 4 regions. Over the whole country, support per FTE dropped in constant dollars. Given that materials expenditures are a significant part of total library expenditures and are central to the mission of the library, the real picture is more pessimistic.
Table 3: Total Library Expenditure per FTE student by Region
Region |
1998-99 Avge. Total Library Expenditure ($/FTE) |
1994-95 Avge. Total Library Expenditure ($/FTE) |
1998-99 Avge Library Expenditure in 1994 dollars-using CPI ($/FTE) |
Atlantic |
$905.71 |
$846.63 |
$850.46 |
Québec |
$642.97 |
$702.00 |
$603.75 |
Ontario |
$932.62 |
$855.63 |
$875.73 |
Prairies |
$843.73 |
$789.92 |
$792.62 |
BC |
$1,008.09 |
$898.96 |
$946.60 |
All CARL Universities |
$844.61 |
$805.84 |
$793.09 |
Figure 2 summarizes these numbers in graphical form.
1.4 Salaries
Table 4: Proportion of Operating Budgets Spent on Salaries by Region
Region |
1994-95 % of Operating Budget Spent on Salaries |
1998-99 % of Operating Budget Spent on Salaries |
BC |
56.6 |
51.9 |
Prairies |
58.7 |
53.5 |
Ontario |
60.6 |
53.3 |
Québec |
61.7 |
58.9 |
Atlantic |
59.6 |
53.6 |
Table 4 shows salaries as a percentage of total library operating budgets by region. Clearly, libraries have reacted to constrained resources by cutting staff and/or restraining salaries.
The numbers of staff employed reflects this trend. In all regions, the average number of professional librarians employed by CARL libraries dropped over the five-year period – in some cases quite dramatically (See Figure 3). The same was true of other full-time staff. At the same time the number of casual staff rose in all regions except UQAM. Only Manitoba, Dalhousie and Toronto reported small increases in total FTE staff over the five-year period.
CISTI bucked the national trend by reporting a significant increase in staff over the five- year period, but this probably partly reflects the fact that CISTI administratively absorbed the formerly separate National Research Council (NRC) Research Press during this period and thus incorporated additional staff. In contrast the NLC went from 504 FTE’s to 416 over the period, a drop of 17.5%.
Despite expenditure constraints, average salaries continued to increase, albeit slowly. Year over year increases in mean professional salaries averaged 1.5% nationwide between 1998-99 and 1999-2000, bringing the average salary up to $60,830. As in the past, salaries are highest in Ontario and lowest in Québec. As has been the case for a number of years, Carleton has the highest average professional salaries – some $7,000 over the next highest – New Brunswick. On the average, CARL professional librarians have 19.5 years of professional experience and have been with their present library for almost 16 years. These figures have not changed significantly in the last five years, suggesting a fairly static workforce.
The average number of administrative librarians has dropped by 2.2/library over the last five years, suggestive of some slight administrative consolidation. This group has an average of 22.3 years of professional experience and has been with the same employer for an average of 16.7 years. They currently make an average of just over $70,000 and, in this category, Carleton has surrendered its long-standing lead to York, where administrative librarians averaged $84,764.
Although data are incomplete, the average salary of "other professional staff" lag a long way behind their librarian colleagues and are closer to average administrative salaries paid in other areas of Canadian universities.
2.0 COLLECTIONS
Since no library weeded more volumes than it added in 1998-99, the collections of all CARL libraries continued to grow.
The number of monographs purchased in 1998-99 compared to the year before dropped in 15 of the 20 CARL libraries reporting this statistic – a reflection of the expenditure squeeze reported above. (In 1997-98 this number had dropped in 10 of the 20 libraries reporting.)
The number of current serial titles (all formats) purchased in 1998-99 compared to the previous year was less in 11 out of 21 libraries reporting. However print/microform serial titles purchased increased in only 4 out of 19 libraries reporting.
The difference is accounted for by the rapid increase in the number of electronic titles purchased. Of those libraries reporting electronic serial purchases, the number of titles varied from a low of 52 at Carleton and McMaster to a whopping 9,206 at Toronto.
Figure 4 shows graphically the dramatic rise in electronic serial subscriptions at the 17 CARL libraries that reported three years of data. At these 17 libraries alone there were more than 25,000 serial titles being purchased in 1998/99. Altogether CARL reported the purchase of over 37,000 electronic serial titles in 1998/99. This does not include the significant numbers of electronic serial titles received free of charge and counted as being included in the member libraries’ collections. Only one library, Carleton, reported a decline in the number of electronic subscriptions purchased over the three years of comparable statistics.
The digital library has clearly arrived and is starting to make inroads on the printed collection.
Needless to say, the large drop in print subscriptions being purchased by most CARL libraries is the result of ongoing cancellations. In 1998-99, except for 5 where data was unavailable, all CARL libraries cancelled serial titles. Numbers of cancellations varied from only 22 at Waterloo to 1,355 at Laval. The annual value of cancellations ranged from $36,690 at the NLC to $1,102,774 at Laval.
This large-scale cancellation of serials is a continuing trend. In 1997-98 cancellations were running at about the same level with a low of 24 cancellations at Queen’s and a high of 2,133 at Saskatchewan. Again the annual value of cancellations ranged from $9,000 to $886,544.
In 1996-97 cancellations ranged from 43 titles at Saskatchewan to 1,763 at the NLC. The annual value of cancellations ranged from $6,700 to $447,148.
No library has been immune from cancelling serial subscriptions and the cumulative effect is a significant decline in holdings. Two of the worst hit libraries over the past three years are Laval with 2,958 cancellations and Saskatchewan with 2,648.
3.0 INTER-LIBRARY ACTIVITY
3.1 Lending
In 1998-99 the largest lender of original documents ("returnables") in the country was Alberta (78,699 loans) followed by the NLC (29,532 loans) and McMaster (19,875). Surprisingly Toronto, by far the largest library, was only fifth. In terms of copies provided, CISTI’s activity dwarfed all others with 466,674 requests filled. British Columbia (28,726), Laval (17,663) and Memorial (17,547) are the next biggest suppliers. Alberta did not supply data, but historically, it has also been the leading CARL university supplier of copies.
Over the three years of comparable statistics overall CARL lending/copying activity has increased steadily, although there have been fluctuations in the output of some of the bigger suppliers – presumably because of active measures to suppress demand.
3.2 Borrowing
In 1998-99 the largest borrower of original documents was Alberta (35,526 loans), followed by Guelph (29,483) and McMaster (11,913). CARL libraries were roughly evenly split between net lenders and net borrowers of originals. On the copy side, the biggest borrower was Guelph (31,654), followed by Calgary (29,126) and BC (24,887). Alberta did not report. The volume of copies "borrowed" has increased at a rapid rate over the last 3 years, with all sources of supply showing marked increases.
Because of gaps in the data it is difficult to do a detailed analysis, but taking figures from the Ontario libraries (excluding Toronto – no data) and Québec (excluding Sherbrooke – no data) the increase in photocopy traffic can be seen as being typical of the rest of the country. In 1996-97 these 14 libraries borrowed 78,173 copies. In 1997-98 this figure rose to 79,376, but in 1998-99 it rose to 132,539. By comparison the number of copies provided by these same 14 libraries was 119,254 in 1996-97, 120,687 in 1997-98 and 127,729 in 1998-99.
It would appear that the ongoing cancellations of serials are finally beginning to be felt in that incoming photocopies outnumbered outgoing copies in 1998-99. This traffic in photocopies is shown graphically in Figure 5.
Clearly, the inter-library lending department is one of the few growth areas in CARL libraries. The major beneficiaries of this increased importation of photocopies appear to be commercial service providers and CISTI.
3.3 Consortium Activities
In 1998-99 all CARL university members reported participating in group purchasing and reciprocal borrowing arrangements, 9 participated in shared storage, 23 in cooperative collection development and 23 in sharing information technology.
There has been a slight increase in participation in these joint activities over the three years of comparable statistics. At the same time the average number of consortia each library was a member of has also increased.
4.0 EMERGING SERVICES
Electronic publishing: In 1996-97, 6 universities plus CISTI stated that they were involved in electronic publishing of textual materials. The following year the number was 7 (not including CISTI, which through the NRC Research Press was actually the largest electronic scholarly publisher in Canada). In 1998-99 the number was 9, while CISTI still did not claim to be in the business.
Quantitative data analysis (data libraries): In 1996-97, 16 universities were involved. In 1997-98, 18 were involved, a number that remained constant in 1998-99. This is clearly a consequence of the "Data Liberation" initiative that started in 1996 and has now spread to the majority of CARL libraries.
Digital imaging: 9 universities offered this service in 1996-97; 8 in 1997-98 and 11 in 1998-99.
Student fee for service: 4 universities had some fees in 1996-97; 3 the next year and 4 the next. This does not appear to be a major trend.
RECON: There seems to be some confusion over this question. In 1996-97 only 8 universities plus CISTI reported that their RECON projects were incomplete. The following year, 10 universities plus CISTI and NLC stated the same thing. In 1998-99, 15 universities plus CISTI and NLC claimed an incomplete RECON project. RECON appears to be the type of project where the more is done, the more remains!
Web access from library workstations: All CARL members reported offering Web access from workstations in the library. In general the number of terminals appears to be growing, although in some cases the numbers reported have decreased. In 1998-99, the number of workstations ranged from 1,493 at Toronto to a single one at the NLC. It is probable that some libraries are not reporting workstations located in the library but administered by other university services, as is the case at Carleton.
Other network services: In 1998-99, all universities offered access to library resources from computer labs; all but 3 offered access to web resources pertinent to the curriculum from the library web page; all but 4 offered a campus-wide information system; all but 10 a File Transfer Protocol (FTP) server; only 7 maintained a gopher server and all offered a web server. All but 4 universities had an electronic classroom or laboratory within the library.
Online Public Access Catalogue (OPAC) workstations: The number of public workstations in the library that provide OPAC access has increased steadily over the three years. In 1998-99 at the universities this number ranged from 1,894 at Toronto to 26 at Windsor. 10 libraries reported having more than 200 such public workstations.
Library systems: In 1998-99, all CARL members except Carleton and UQAM used a commercial OPAC system. DRA had 6 customers; Innovative had 5; Endeavor (Elsevier) had 5; Ameritech/NOTIS had 5 (various versions); SIRSI had 4 and Geac had 2.
Twelve CARL members reported their OPAC as being part of an integrated network system. All but three OPAC systems (York plus the two libraries maintaining home-grown systems) included a Z39.50 compliant server for standardized client-server access. Two years earlier 5 libraries did not have such software.
All but 6 CARL members offer access to other libraries’ holdings through their OPAC. Two years previously the corresponding number was 9, reflecting the increasing inter-linking of library systems.
5.0 HARDWARE, SOFTWARE AND ELECTRONIC INFORMATION EXPENDITURES
5.1 Hardware Expenditures
Electronic hardware expenditures are becoming an increasingly large component of CARL library budgets. In 1998-99 these expenditures ranged from $1.9 million at the NLC and $1.5 million at Toronto to $21 thousand at Manitoba. The average expenditure on hardware was $420,350. This compares with 1996-97 average expenditures of $353,757 (Data for the 23 libraries reporting in both years)
5.2 Software Expenditures
On the software side, average expenditures have gone down between 1996 and 1998 for those libraries reporting in both years. Average expenditures fell from $485,618 to $149,896. (23 libraries reporting)
5.3 Electronic Information Expenditures
For the 21 libraries that reported data for the three years 1996, 1997 and 1998, the average expenditure on electronic information rose from $319,844 to $523,858 a year later and then fell to $520,050. The range of expenditures in 1998-99 was considerable: from $1.87 million at Toronto to $272 thousand at Memorial.
6.0 CIRCULATION ACTIVITY
Comparing circulation statistics is a dubious activity, as the numbers reported are so dependent on local borrowing policies and definitions. Presumably, however, it is valid to compare activity reported at a given library over time. Circulation data for 1996-97 compared to that for 1997-98 shows that circulation is down in 15 libraries and up in 12 (2 did not report). It does not appear that these changes in circulation patterns can be explained only in terms of changing enrolment. Rather there seems to be an underlying drop in the average circulation per user in many CARL libraries. The most likely explanation would be a shift to direct online access to electronic information sources, both provided through the library and from other sources.
For example at Toronto circulation transactions dropped 14.3% over the three years of comparable data. FTE student enrolment went down less than 1% during the same period. In-house use of the collection also dropped, but only by 3.0%. In the same period Toronto invested heavily in electronic materials and workstations. It would appear that the effect, as would be expected, is to displace some demand from the print collection to the electronic collection. The same effect is observed at Canada’s second biggest library. At UBC, circulation transactions dropped by 21.6% over the three years, while enrolment went up 3.4%. Again UBC has invested heavily in electronic materials and access.
Figures for in-house use are spotty and of questionable reliability, but where data is reported the trend is more often down than up. Data reported for reserve circulation transactions is also spotty but, for those libraries reporting for both 1996 and 1998, traffic is down in all but 2 libraries. In some cases the drop in traffic reserve circulations is as much as one third. This would appear to be another manifestation of the impact of the Web. More and more reserve readings are being distributed, in electronic form.
7.0 SERVICE HOURS, INSTRUCTION AND REFERENCE ACTIVITY
In 1998-99 the average number of staffed library service points was 17.8. The number of weekly public service hours offered by CARL university libraries was 90.5. In 1996-97 the average number of staffed service points was slightly greater at 18.6 and the average hours of service per week was slightly less at 86.9.
Reference transactions are even more unreliable and difficult to compare than circulation transactions. This being said comparing 1996-97 to two years later, transaction volume declined in 21 out of 29 CARL libraries. Once again one can surmise, without much evidence, that the Web is supplanting some traditional informational reference activity.
The CARL/ABRC1998-99 Statistics/Statistiques is available for $50 (plus shipping and handling) from:
Canadian Association of Research Libraries/
Association des bibliothèques de recherche du Canada
Room 139/ Pièce 139
Morisset Hall/ Pavillon Morisset
University of Ottawa/Université d`Ottawa
65 University Street/rue Université
Ottawa ON KIN 9A5
Tel: (613) 562-5800 x3652
Fax: (613) 562-5195
Email: carladm@uottawa.ca
