You are on page 1of 15

Measuring Service Quality at Yale University’s

Libraries
by Danuta A. Nitecki and Peter Hernon

This article examines a new


approach to measuring service
quality, one that produces
findings useful for local
F or years, researchers in library and
information science (LIS) have ex-
amined information needs, user
wants, and user perceptions about the
value of library services. They have also
survey instrument based on the Gaps
Model of Service Quality. The model
summarizes a set of five gaps showing the
discrepancy between:
● Customers’ expectations and manage-
planning and decision making. looked at an elusive concept—quality—in
terms of collections (size, titles held, and ment’s perceptions of these expecta-
It also differentiates between breadth of subject coverage) and the ef- tions (Gap 1);
service quality and satisfaction fectiveness (extent to which goals and ob- ● Management’s perceptions of custom-
and points the direction for jectives are set and met) of library ser- ers’ expectations and service quality
vices. In recent years, LIS researchers specifications (Gap 2);
further research. have drawn on marketing and other liter- ● Service quality specifications and ac-
atures to focus attention on expectations
and an alternative view of quality, one tual service delivery (Gap 3);
representing the user’s or customer’s per- ● Actual service delivery and what is
spective on the services used. Those re- communicated to customers about it
searchers who have examined quality (Gap 4); and
from that perspective concur with their ● Customers’ expected services and per-
peers in marketing that “only customers ceived service delivered (Gap 5).6
judge quality; all other judgments are es-
sentially irrelevant.”1 As Valarie A. The first four gaps are the major con-
Zeithaml, A. Parasuraman, and Leonard tributors to the service quality gap that
L. Berry emphasized, “service-quality customers may perceive. The fifth gap is
perceptions stem from how well a pro- the basis of a customer-oriented definition
vider performs vis-à-vis customers’ ex- of service quality; it is the discrepancy
pectations about how the provider should between customers’ expectations for ex-
perform.”2 As a result, the research has cellence and their perceptions of actual
tended to define service quality in terms service delivered. This discrepancy is the
of meeting or exceeding customer expec- conceptual basis for the SERVQUAL in-
tations,3 or, more precisely, as the differ- strument.7 By applying the instrument,
ence— or gap— between customer per- libraries gain insights into customers’
ceptions and expectations of service. conceptualization of what a service
Recently, Peter Hernon and John R. should deliver and how well that service
Whitman argued that it is possible to ex- meets idealized expectations.
amine expectations from two coequal and Another approach, one also built
probably interrelated concepts, service around the Gaps Model, centers around
quality and satisfaction.4 Service quality the work of Hernon and his colleagues in
Danuta A. Nitecki is Associate University deals with those expectations that the li- the United States and New Zealand.8
Librarian, Yale University Libraries, Box brary chooses to meet, and satisfaction is They, likewise, look at the service nature
208240, New Haven, Connecticut 06520- more of an emotional and subjective re- of libraries but view service quality in the
8240 ⬍danuta.nitecki@yale.edu⬎; Peter action to a time-limited event or the cu- context of library planning and decision
Hernon, Professor, Simmons College, mulative experiences that a customer has making. They have suggested a pool of
Graduate School of Library and Information with a service provider.5 more than 100 candidate service attributes
Science, 300 The Fenway, Boston, Within LIS, different approaches to that staff could review and from which
Massachusetts measure service quality are emerging. they could select a subset potentially hav-
⬍peter.hernon@simmons.edu⬎. One centers on the use of SERVQUAL, a ing the greatest relevance to their library.

The Journal of Academic Librarianship, Volume 26, Number 4, pages 259 –273 July 2000 259
Thus, these researchers maintain that any tended to find “reliability” as the most [they do] not imply that the SERVQUAL
subset of attributes that staff members important dimension and “tangibles” as five dimensions are irrelevant to the aca-
select varies from institution to institution the least important one. She observed that demic library setting. They do, however,
suggest that interpretations based on the cal-
because different libraries are unlikely to the instrument’s designers had a similar
culated dimensions should be made with
be willing to provide the resources neces- finding when they studied various service caution. Additional research in other aca-
sary to meet or exceed exactly the same industries in the profit sector. She also demic libraries is needed to confirm the ap-
set of expectations. noted that there is little overlap among the plicability of these five dimensions or to
five dimensions and differences regarding identify a different model for academic li-
SERVQUAL which dimensions are, indeed, important brary services.17
SERVQUAL, first introduced in 1988 in when measured by the grouped state-
marketing, consists of 22 pairs of state- ments.10 Reviewing eight published ap- Even though the length of SERV-
ments, the first of which measure the ex- plication studies of SERVQUAL in li- QUAL, and the repetition of the set of 22
pectations of a service provider’s custom- brary settings, Nitecki observes that statements (as part of both ideal and ac-
ers by asking each respondent to rate, on tual service expectations), make the data
a seven-point scale, how essential each none of the library settings. . . [provides ev- collection instrument somewhat cumber-
item is for an excellent service provider to idence that] the five dimension patterns pro- some to use, return rates have varied from
deliver. The second set of 22 identical posed by the SERVQUAL designers applies 27% to 76%18 and the reliability of test
to. . . [those] settings. Except for the Mary- scores has been demonstrated. Library
statements ascertains the respondent’s
land study, all data agree that without con- staff, however, cannot extensively alter
perceptions of the level of service given straint, a five factor solution can be estab-
by the institution or organization exam- the statements of expectations within the
lished, but the items do not load on the same
ined. For each pair of statements, the dif- standard SERVQUAL design to address
factors as developed by the SERVQUAL
ference between the ranked perception designers. There is no repetition of associ-
local circumstances and preferences with-
and the ranked expectation is calculated; ated patterns among the library studies, ex- out challenging the designer’s assertion
the average of these gap scores is the cept that consistently the first four statements that the statements fully reflect the dimen-
SERVQUAL overall quality score. The load together (the Tangible group) and state- sions.
instrument’s designers—Zeithaml, Para- ments 18, 20 –22 (Empathy). The Maryland
suraman, and Berry— claim that each set study reached a four-factor solution.11 THE PRESENT STUDY
of 22 statements encompasses five inter- Problem Statement
related dimensions that customers most Syed Saad Andaleeb and Patience L.
value when, regardless of the service pro- Simmonds,12 as well as Emin Babakus SERVQUAL is the data collection in-
vider used, they evaluate service quality: and Gregory W. Boller,13 questioned the strument most frequently used to measure
extent to which SERVQUAL captures all service quality and to convey how cus-
● Tangibles (physical facilities, equip- five dimensions consistently. Andaleeb tomers judge the quality of service in
ment, and the appearance of person- and Simmonds found that a dimension, many service industries. Although its ap-
nel); demeanor, was an important aspect of li- peal to libraries is growing, SERV-
● Reliability (ability to perform the brary service and that it “is a rough com- QUAL’s standardized statement of ser-
promised service dependably and ac- bination of two SERVQUAL dimensions, vice attributes as the basis for judging
curately); empathy and assurance, and the helpful- service quality in libraries limits its appli-
ness criterion normally associated with cability for improving specific local ser-
● Responsiveness (willingness to help responsiveness.”14 Colleen Cook and vices. On the other hand, the approach
customers and provide prompt ser- Bruce Thompson surmise that taken by Hernon and his colleagues19 en-
vice); ables a library to identify those attributes
● Assurance (knowledge and courtesy of It is possible that the Demeanor factor. . . of greatest local importance for service
employees and their ability to inspire speaks to a similar concept. . . identified in improvement. No studies have attempted
trust and confidence); and the factor, “Affect of Service Experience,” to combine both approaches and to deter-
which is a close, but not exact amalgamation
● Empathy (caring, individualized atten- mine the extent to which the resulting
of SERVQUAL’s responsiveness, assurance,
tion the firm provides its customers).9 and empathy dimensions, and that these ele-
instrument provides insights useful for lo-
ments may constitute one rather than three cal library planning and decision making.
The SERVQUAL designers contended factors. Other studies in retailing and in This study customized the instrument
that the 22 statements group, as factor banking, motor vehicle, brokerage, electrical for local use, while accepting the Gaps
loads, onto these five dimensions and, appliance and life insurance services indus- Model of Service Quality, which is fun-
thus, define the dimensions. They further tries have yielded similar results.15 damental to the design of the SERV-
assert that the average scores of the state- QUAL instrument. The goal of the instru-
ments grouped around each dimension They suggested that there may be three ment and approach that this study
correspond to the dimension’s service dimensions in libraries, tangibles, reliabil- explored was to target actual service ele-
quality gap score. Another key part of the ity or service efficiency, and affect of ments for improvement, to weigh the
instrument is a direct request of customers service, and that there is a need for further evaluation of service elements relative to
to indicate the relative importance that research to explore the dimensions “that the importance that customers place on
they attach to these dimensions by a 100- may underlie quality service as a con- them, and to encourage the allocation of
point allocation. struct in the research library setting.”16 resources for meeting those expectations
Danuta A. Nitecki reported that studies Nitecki concludes the following about the that a library and its customers deem im-
using SERVQUAL in libraries have findings of past studies: portant. Nitecki added that:

260 The Journal of Academic Librarianship


agreed to collaborate in the investigation quality requires input from current read-
“The goal of the instrument of the study objectives. As a first step, ers, we rejected any technique that inves-
and approach that this study Peter Hernon spoke before library staff on tigated the entire population (or drew a
customer service and service quality. The probability sample) of students, staff, and
explored was to target actual Service Quality Improvement Council, faculty. Some of these individuals might
service elements for composed of staff throughout the library not use the libraries. Completing in-house
improvement, to weigh the system, then reviewed the “Aspects of surveys at Yale University is uncommon,
evaluation of service elements Service Quality” and “Sample Statements and we were, therefore, concerned that
on Service Quality Expectations”22 to de- some readers might resent being ap-
relative to the importance that velop a set of core service attributes for proached within the library and asked to
customers place on them, and library readers to assess. Once that list participate. Telephone and in-person in-
to encourage the allocation of had been developed, it was shared with terviewing were also rejected as too time-
resources for meeting those various groups of readers, predominately consuming and costly to execute. An e-
students, which suggested some alterna- mail survey was also not pursued because
expectations that a library and tive attributes and changes in the wording circulation records do not record e-mail
its customers deem important.” for different statements of expectations. addresses; thus, it would have been very
The university librarian and the associate time-consuming to ascertain those ad-
Academic library staff and managers need to university librarian (for reader services) dresses. As a result, we decided to risk a
better understand what influences users’ reviewed the emerging list, recommend- low return rate and a possible self-se-
judgments of service quality, what is ex- ing some alternative wording and at- lected sample. We distributed a printed
pected from specific services, and what im- tributes, and concluded that the list of test questionnaire through the U.S. mail; re-
provements in service design and delivery service attributes was meaningful for spondents were asked to complete and
are effective. They need to understand their planning and decision-making purposes. mail it back to the library.
users not merely as recipients of services Finally, we reviewed the draft document From the population of those readers
offered, but as partners in the development and placed the various attributes in the who had borrowed a book within the past
and implementation of services to make
form of a questionnaire that, basically, year, a random sample of 500 participants
higher education and research experiences
more successful from the customer’s per-
adhered to the SERVQUAL format. We was selected in the fall 1999. The partic-
spective.20 reviewed the format of the instrument and ipants were mailed a copy of the question-
the sequencing and wording of statements naire, together with a cover letter and a
Objectives and also asked the university librarian for return envelope (with postage provided).
The study objectives were to: any final comments. Based on his re- The identity of participants was kept con-
sponses, we rewrote some statements and fidential and retained only for the purpose
● Develop and test an approach for con- verified that each question on the survey of tracking who returned and did not re-
verting SERVQUAL into an instru- had value for library planning and deci- turn completed questionnaires. Later that
ment that reflects the expectations of a sion making. semester, non-respondents were sent a
library and its customers; The final questionnaire (see Appendix second copy of the form, with a new
● Determine the feasibility of libraries A), is divided into three sections. The first cover letter, and invited to participate in
pursuing that approach; and one combines two parts of SERVQUAL the study.

and explores both reader expectations and A total of 226 faculty members, stu-
Compare study findings to other uses delivery perceptions regarding 40 state- dents, university staff, and other (e.g., vis-
of SERVQUAL within LIS. ments. Respondents were also given an iting scholars) responded, for a return rate
Unlike the work of Colleen Cook and opportunity to identify any other expecta- of 45.2%. It is interesting to note that one
Bruce Thompson,21 our goal was neither tions. The next section contains state- faculty member wrote a letter in which he
to produce an instrument that can neces- ments of the same five dimensions found explained that, although he did not want
sarily be applied across institutions nor to in SERVQUAL and seeks readers’ per- to complete the survey, he believes the
use the resulting data to make compari- ceptions of the relative importance of library provides “excellent” service.
sons among so-called peer institutions. each dimension to their evaluation of ser- Moreover, he inquired about how to make
Rather, our work centered on developing vice quality; thus, it is possible to make a financial contribution to the library’s
and testing an approach by which individ- some general comparisons to related re- collection!
ual institutions can set and assess those search. The final section contains an over- Survey responses were entered in an
expectations that have local value. The all expectations question and seeks (1) Excel file and the data were transferred to
question not addressed in our research is information about service provided at the Statistical Package for the Social Sci-
“Does a set of quantifiable core service various Yale University libraries, and (2) ences (SPSS for the PC version 10.0) for
attributes or expectations exist across in- descriptive information about the respon- data analysis. All of the data were verified
stitutions and user groups?” That question dents. The survey was printed on 16“ ⫻ to ensure the accuracy of data entry, and
is premature for us to probe at this time. 11” paper and folded in half; thus, pre- significance testing was set at p ⫽ .05.
senting on one piece of paper a page for
Procedures the cover letter and three pages for the FINDINGS
The setting for this study was Yale survey. As a result, it was easy for respon- An examination of completed survey
University, where the library staff refer to dents to view the entire questionnaire at a forms indicated that respondents did not
users or customers as readers (hereafter quick glance. indiscriminately mark the same number
this word will be used). In 1998, we Because an examination of service for each statement. Rather, they appeared

July 2000 261


to have read and thought about each state- Figure 1
ment before answering it. Further evi- Depiction of Quadrants
dence of the care that respondents took in
completion of the survey is that 171
(76%) wrote brief comments on the form.
Respondents
Of the 226 respondents, 224 identified
their status as faculty, students, staff, or
other. Overwhelmingly, they were gradu-
ate or undergraduate students (183 or
81.7%); 23 (10.3%) were faculty, 14
(6.2%) were staff, and 4 (1.8%) classified
themselves as “other.” To address whether
the respondents comprised a self-selected
sample, we compared the status of the sam-
ple to that of the respondents. There was no
significant difference between both groups
(␹2 3.213, df ⫽ 3, p ⬎ .05). The respon-
dents consequently do not represent a self-
selected sample; findings can be generaliz-
able to the population of Yale readers who
have recently borrowed material.
Only one respondent (0.04%) marked
nonuse of the libraries during this school
term; 186 (82.7%) used the libraries at
least once a week. Among the respon-
dents, 42 (18.7%) characterize their li-
brary use as daily, and 95 (42.2%) indi-
cated their use was several times per
week.
Which of the 40 Service Attributes
Quadrant analysis, which is frequently expectations, these service features merit
Are Very Important to the Customers
used in marketing, is a graphic correlation improvement. Any attributes present in
and for the Library to Meet?
technique that produces data easy to visu- Quadrant 3 are relatively unimportant to
Respondents, on average, ranked the alize.23 The technique plots data about the readers, but readers associate those
importance of the 40 attributes that con- service attributes into four quadrants de- attributes with library service. The staff
tributed to the library’s quality of service fined by two dimensions: one reflects the might refocus the service so that its image
between 5.1 and 6.8 on the seven-point importance that customers give service matches the attributes shown in Quad-
scale. This relatively high range of rank- attributes, while the other indicates the
ings suggests that all of the factors se- extent to which customers think a partic-
lected by the managers as important were ular service has the attributes. Figure 1
so to the readers. The perceptions among illustrates the implications assigned to at- Figure 2
readers of how well the Yale University tributes falling within the four quadrants, The 40 Service Attributes Studied
libraries deliver each service attribute when the first dimension is plotted along
ranged, on average, between 4.0 and 5.9 the horizontal axis as the ideal expecta-
on the seven-point scale. In absolute tion for excellent service quality, and the
terms, these rankings of performance are second dimension is plotted along the ver-
all above the scale’s midpoint but indicate tical axis as the library’s performance.
little about the relative importance of an Those attributes falling into Quadrant
attribute to each score. However, when 1 are very important to the readers, and
considered in relation to the expectations these users perceive the library as pos-
placed on each factor, the resulting corre- sessing them or as performing well in
lations offer insights about where manag- their delivery. These are attributes that the
ers should focus attention to improve the library should retain in any reconfigura-
quality of service provided. The relation- tion of its services and that the library
ship between the two measures— expec- should not see a decline in its ability to
tations for excellence and perceptions of meet. Attributes falling into Quadrant 2
service delivered— can be presented in are also most important to the readers but
one of two ways, visually through quad- are not perceived as being prominent fea-
rant analysis and numerically as gap tures of a library service. Within a library
scores. culture that strives to respond to reader

262 The Journal of Academic Librarianship


rants 1; alternatively, library staff might ● Materials I requested to be paged ● The online catalog is easily accessible
want to revisit resources allocated to pro- come within the time frame quoted by from outside the library building (gap
viding the service attributes in Quadrant 3 the library staff (gap of ⫺1.5); of ⫺0.74);
and review the relative expense of provid- ● When I request materials, I am told ● The staff are courteous and polite (gap
ing a less valued activity. Quadrant 4 in- how long they will take to arrive through of ⫺0.80);
cludes attributes that are neither valued by interlibrary loan (gap of ⫺1.5);
readers nor performed well by the library. ● The library Web site enables me to log
● The online catalog has easy-to-follow
These attributes might be ignored and re- on easily (gap of ⫺0.87);
sources reallocated. instructions (gap of ⫺1.6);
● The staff are expert in finding general
Quadrant analysis was used to display ● Computer printers in good working or-
the data gathered (see Figure 2). Eleven information (gap of ⫺0.90); and
der are available when I need them
attributes were very important to readers, (gap of ⫺1.9); ● Library materials meet my course/re-
and the library performed them well ● Photocopiers in good working order search needs (gap of ⫺0.92).
(Quadrant 1), and three attributes were of
are available when I need them (gap of Other Expectations
relatively low importance and low perfor-
⫺1.9);
mance (Quadrant 4). These 14 attributes
carry less urgency to review for improve- ● Material are in their proper places on When asked if there were other expec-
ment in performance (see Table 1). Read- the shelves (gap of ⫺1.9); tations that they considered important, 75
ers ranked 26 attributes as important for respondents (33.2%) responded in the af-
● Directional signs are clear and helpful
service excellence but they feel that these firmative and listed one service attribute;
(gap of ⫺2.0) 32 (14.2%) identified a second attribute.
attributes are less evident in the library’s
delivery of service quality (Quadrant 2),
● Materials are reshelved promptly (gap The attributes suggested tended to focus
and they did not identify any factors as of ⫺2.1); on the comprehensiveness of collections,
unimportant but within the library’s per- ● It is easy to find out, in advance, when remote access to those collections, accu-
ceived service strengths (Quadrant 3). the library is open (gap of ⫺2.1); and racy of the check-in process for books
once readers returned them to the library,
● The online catalog is an accurate and the amount of fines (e.g., “fines for
source of information about all mate- overdue books are reasonable”). There
“Readers ranked 26 attributes rials held by the library (gap of ⫺2.2). was also some repetition with the at-
as important for service tributes previously discussed, such as
It is important to remember that the
those covering library hours, equipment
excellence but they feel that judgments about the importance of the
in good working order, and the need to
these attributes are less evident attributes and the perceptions of services
reshelve material in a timely manner. One
delivered are relative and are merely in-
in the library’s delivery of dicators of where priorities might be
respondent wanted the library to give
service quality (Quadrant 2), more consideration to the “ergonometric
placed for improvement efforts. In some
character of the library environment.”
and they did not identify any instances, more exploration may be
factors as unimportant but needed to understand more fully the na- Percent Allocation to the SERVQUAL
ture of reader’s expectations and how ser- Dimensions
within the library’s perceived vices can be improved.
service strengths (Quadrant Finally, 12 statements had a quality Section B of the survey (Appendix A)
3).” gap score (difference between the ideal listed the five dimensions of SERVQUAL
and actual mean scores) of less than 1.00: applicable to libraries and the services
they offer. The investigators queried re-
● The library Web site is attractive (gap
A second technique to present survey spondents about the importance of each
of ⫺0.1); dimension to them when they evaluate the
findings is to calculate gap scores and
review the numerical relationship among ● Library materials encompass curricu- quality of library service. Respondents
the attributes. Based on this analysis, the lum-supporting videos and films (gap were asked to allocate 100 points among
library did not surpass reader expectations of ⫺0.8); the five dimensions according to how im-
on any service attribute, but it did come ●
portant each one was to them. Table 2
Computer dedicated only for online
close to meeting most expectations. The displays the results. “Reliability” received
catalog use in good working order is
the largest percentage allocation and “em-
difference between expectations and per- available when I need it (gap of ⫺0.9);
ceptions for 17 factors fell within one pathy” the least.
● The staff provide assistance to help me When asked if there was anything else
point of meeting expectations (when the
evaluate information I find (gap of not included in the five dimensions that
gap score was rounded to the full digit),
⫺0.9); they find important in evaluating the qual-
whereas the gap for 11 factors fell beyond
one point. The following service at- ● There is a paging service for retrieving ity of service received, 63 (27.9%) re-
tributes should command further explora- materials (gap of ⫺0.9); sponded in the affirmative. Their re-
tion because they are the furthest from sponses tended to focus on the adequacy
● The online catalog indicates the num-
meeting reader expectations: or comprehensiveness of collections, li-
ber of copies available (gap of ⫺0.56); brary hours, the return of books, the fine
● The library Web site enables me to in- ● The staff are friendly and easy to talk policy, and the ease in finding materials
teract with library staff (gap of ⫺1.5); to (gap of ⫺0.73); they sought.

July 2000 263


Table 1
Summary of Quadrant Analysis Findings for the 40 Expectations Probed in the Survey1
Quadrant 1 (Important Factors to Retain—Maintain High Performance)
1. Library materials meet my course/research needs.
2. The library Web site enables me to log on whenever I want.
3. The library Web site enables me to log on easily.
4. The staff are expert in finding general information.
5. It is easy to find out, in advance, when the library is open.
6. The staff are approachable and welcoming.
7. The staff are courteous and polite.
8. The library Web site enables me to access a variety of electronic resources.
9. The online catalog indicates the number of copies available.
10. The online catalog is easily accessible from outside the library building.
11. The staff are friendly and easy to talk to.
Quadrant 2 (Important Factors to Improve Performance)
1. The online catalog displays information that is clear and easy to understand.
2. The library Web site is easy to navigate.
3. The staff are available when I need them.
4. The online catalog has easy-to-follow instructions.
5. Computer research workstations (e.g., for access to the Web) in good working order are available when I need them.
6. The staff provide assistance to help me learn how to find information.
7. When I request materials, I am told how long they will take to arrive from restricted collections.
8. It is easy to browse print collections.
9. Materials I requested to be paged come wtihin the time frame quoted.
10. The library Web site includes online request forms (reference and interlibrary loan).
11. The staff provide assistance to help me retrieve resources I need.
12. The staff provide assistance to help me identify resources I need.
13. Computers dedicated only for online catalog use, in good working order, are available when I need them.
14. The staff are expert in the literature of my discipline.
15. Material are in their proper places on the shelves.
16. It is easy to find where materials are located in the building.
17. Materials are reshelved promptly.
18. Directional signs are clear and helpful.
19. Photocopiers in good working order are available when I need them.
20. The online catalog is an accurate source of information about all materials held.
21. When I request materials, I am told how long they will take to arrive through interlibrary loan.
22. Computer printers in good working order are available when I need them.
23. Microform/fiche readers in good working order are available when I need them.
24. Microform/fiche reader/printers in good working order are available when I need them.
25. There is a paging service for retrieving materials.
26. Library materials encompass curriculum-supporting videos and films.
Quadrant 4 (Relatively Unimportant Factors Not to Address Immediately)
1. The library Web site enables me to interact with library staff.
2. The library Web site is attractive.
3. The staff provide assistance to help me evaluate information I find.
1
Note: See Appendix B for an explanation of the quadrants.

264 The Journal of Academic Librarianship


Table 2 Thomas A. Childers and Nancy A. Van
The Five SERVQUAL Dimensions House found that library users show a
marked preference for self-service,24 and
Percent Allocation Mean Peter Hernon and Ellen Altman in their
Dimension of 100 Points (Standard Deviation) Number research into service quality noted a cus-
Tangibles 1
18 17.64 (12.14) 221 tomer preference to be self-reliant.25 In-
2
formation-handling technologies, it
Reliability 32 31.87 (13.75) 222 seems, tend to promote a desire for self-
Responsiveness3 22 21.72 (9.07) 222 service and self-reliance. An examination
Assurance 4
16 16.28 (7.66) 222 of the factors appearing in both Quad-
5
rants 1 and 2 (see Table 1) suggests that
Empathy 12 12.45 (6.91) 221 readers, to some degree, wanted to navi-
Notes: 1
The appearance of the library’s physical facilities, equipment, and communication materials. gate the library on their own. As indicated
2
The library’s ability to perform promised services dependably and accurately. by the statements in Quadrant 1, they
3
The library’s willingness to help readers and provide prompt services.
4
The knowledge and courtesy of the library staff and their ability to inspire trust and confidence. showed a willingness to consult with
5
The caring, individualized attention the library provides to its readers. staff. They regarded the demeanor of the
staff as very important and valued “easy”
access to library materials, both remotely
Overall Expectations Met order to avoid inappropriate fines.” Some via the Web and on-site. As well, some
respondents perceived that, although they statements about the online public access
Respondents were asked to rate, on a
had returned books on time, they were, catalog (OPAC) and the library’s Web
10-point scale ranging from falls short to
nonetheless, charged for being delin- site appeared in the first quadrant. “Easy”
meets and to surpasses, “Overall, to what
quent. They were dissatisfied with the access also frequently appeared among
extent, does the service that Yale Univer-
outcome of the appeals process and, in a the statements in Quadrant 2. Here, Yale
sity libraries provide meet your expecta-
few instances, with the process itself. readers also expressed a desire to have
tions for an excellent library?” The mean
Some respondents wanted the notice
response to this question was 7.11 [n ⫽ sources in their proper places and to have
for overdue materials and for the notifi-
221, SD ⫽ 1.79]. The mean rating on the displaced sources promptly reshelved.
cation of recalled books to be sent via Readers wanted staff to be experts in find-
seven-point scale for how Yale University
e-mail, and some believed that fines were ing general information (Quadrant 1) and
libraries deliver services on the 40 factors
set too high. There was also repeated to be experts in the literature of their
was 5.1224; if converted to a 10-point
mention of one library and the perception disciplines (Quadrant 2). They also val-
scale for comparison, this Yale perfor-
of poor quality of service provided [e.g., ued staff assistance in learning how to
mance mean was 7.32. The correlation
“a good amount of important books are find, identify, and evaluate information
(using the Pearson Product-Moment Cor-
missing or could not be found (even by (Quadrant 2).
relation, r,) between the mean scores of
staff)”], and the “high cost” of photocopy- Because section B of the survey pro-
responses to the question and the perfor-
ing in comparison to what peer institu- vides a direct connection between this
mance factors was positive and signifi-
tions charge or the “high cost and poor
cant (r ⫽ 0.523, p ⬍ .05). Clearly, Yale study and other applications of SERV-
quality” in comparison to what “a stan- QUAL itself, comparing percentage allo-
University libraries, overall, are meeting,
dard copy shop” offers. cation among various studies can provide
but not surpassing, customer expectations.
Some respondents mentioned that they insight (see Table 3). Excepting the study
Compliments had requested a book that was either in reported here, “reliability” is the most im-
Some respondents took the time to processing or on order, but were disap- portant dimension and “tangibles” is the
compliment the library for having high- pointed that they were not notified when least important. This study shows that “re-
quality collections and library staff for the book was placed in the collection. liability” produced the largest percentage
their “excellent” assistance. As one re- Others thought that the online public ac- allocation, but that “empathy” was the
spondent wrote, “I love Yale’s libraries. cess catalog did not always accurately list lowest. It would seem that the five dimen-
They are my pride and joy. Thank you for availability: titles listed as available, they sions merit reconceptualization. De-
working to improve them.” explained, might not be on the shelf. meanor cuts across a couple of dimen-
Some even mentioned the need for better sions, and the five dimensions fail to
Potential Problems signage and lighting in the stack area adequately address the desire to be self-
The question about having other ex- (e.g., “differentiate between emergency supporting or self-reliant.
pectations, together with the questions exits and regular exits”). Similar patterns appeared in compari-
about the comprehensiveness of the five sons with questions posed in numerous
dimensions and about any library depart-
DISCUSSION library applications of SERVQUAL that
ment or service that did not merit a rank- The respondents to this survey were sim- asked respondents to identify the most
ing of “7” on a 10-point scale, produced ilar to those in a number of other studies and least important dimension (see Table
opportunities for respondents to point out applying the SERVQUAL instrument in 4). Except for a study of reference ser-
problems. A number of the written com- academic libraries, namely they tended to vices, all applicants identified reliability
ments focused on the need for the return be students. Typically, in other studies as the single most important service di-
of books to be “easy and secure.” Or, as conducted at Yale, faculty accounted for mension. With the exception of the cur-
one respondent phrased it, an expectation 4 – 6% of the respondents; in this study, rent study, library customers identified
could be “discharging books properly in the percentage was 10.3. tangibles as the least important.

July 2000 265


Table 3
Comparison of Percent Allocation of 100 Points to the Five SERVQUAL Dimensions in Different Studies
Dimensions1 (%)
Study Service Setting N Tangibles Reliability Responsiveness Assurance Empathy
a
Zeithmal et al. Multiple industries 1,936 11 32 22 19 16
(non-library)
Nitecki & Hernonb Academic library 221 18 32 22 16 12
c
Coleman et al. Academic library 198 16 27 24 19 15
Edwards & Browned Academic library 80 9 36 23 17 15
Niteckie Academic library 140 9 39 23 17 13
(ILL)
Steinf Academic library 246 8 41 24 16 11
(ILL)
Hébertg Public library 130 12 35 20 20 14
(ILL)
Niteckie Academic library 95 10 26 25 22 18
(reference)
Niteckih Academic library 90 10 31 23 24 16
(reference)
Niteckie Academic library 101 9 35 24 19 13
(reserve collection)
Whitei Special libraries n/a 12 34 23 18 13
1
Notes: Based on calculated percent point allocation on questions B1 through B5.
a
Valarie A. Zeithaml, A. Parasuraman, & Leonard L. Berry, Delivering Quality Service: Balancing Customer Perceptions and Expectations (New York: The Free Press, 1990).
b
The reference is to the present study.
c
Vicki Coleman, Yi (Daniel) Xiao, Linda Blair, & Bill Chollett, “Toward a TQM Paradigm: Using SERVQUAL to Measure Library Service Quality,” College & Research
Libraries 58 (May 1997): 237–251.
d
Susan Edwards & Mairéad Browne, “Quality in Information Services: Do Users and Librarians Differ in Their Expectations?” Library & Information Science Research 17 (1995):
163–182
e
Danuta A. Nitecki, An Assessment of the Applicability of SERVQUAL Dimensions as Customer-based Crietera for Evaluating Quality of Services in an Academic Library.
Unpublished doctoral dissertation, University of Maryland, 1995.
f
Joan Stein, “Feedback from a Captive Audience: Reflections on the Results of a SERVQUAL Survey of Interlibrary Loan Services at Carnegie Mellon University Libraries,” in
Proceedings of the 2nd Northumbria International Conference on Performance Measurement in Libraries and Information Services (Newcastle upon Tyne, England; Department
of Information & Library Management, University of Northumbria at Newcastle, 1998), pp. 207–222.
g
Françoise Hébert, “Service Quality: An Unobtrusive Investigation of Interlibrary Loan in Large Public Libraries in Canada,” Library & Information Science Research 16 (1994):
3–21.
h
Danuta A. Nitecki, “Assessment of Service Quality in Academic Libraries: Focus on the Applicability of the SERVQUAL,” in Proceedings of the 2nd Northumbria International
Conference on Performance Measurement in Libraries and Information Services (Newcastle upon Tyne, England: Department of Information & Library Management, University
of Northumbria at Newcastle, 1998), pp. 181–196.
i
Marilyn D. White, Measuring Customer Satisfaction and Quality of Services in Special Libraries (unpublished final report to Special Libraries Association, 1994).

Written comments on the question- fewer dead terminals),“ and another was cator of service quality should not be
naire reflected some dissatisfaction with even stronger in making the point: “ap- taken out of context.
the five dimensions adequately covering pearance means nothing—functionality is Clearly, SERVQUAL is not an instru-
library services. For example, some re- everything.” ment for measuring satisfaction, and the ro-
spondents commented that none of the
dimensions addressed the breadth, depth, Overall Expectations Met
and content of collections. Others men- As previously noted, the mean score “SERVQUAL is not an
tioned that they did not see the difference to characterize the extent to which re- instrument for measuring
between “the library’s willingness to help sponding readers’ expectations were satisfaction, and the robust
readers and provide prompt service” and met was 7.11. In general, a library
“the caring, individualized attention the might strive to achieve a mean of at
nature of the SERVQUAL score
library provides to its readers.” They also least 8 to shrink the breadth of the scale measuring the gap between the
saw some overlap between these state- (7–10), conveying more than a sense ideal expectations and actual
ments and “the knowledge and courtesy that service expectations were surpassed service performance is a
of the library staff and their ability to and not merely met.26 However, before
inspire trust and confidence.” One respon- making this characterization and taking meaningful gauge of quality
dent argued that appearance is less impor- action to improve on a mean score of performance for library
tant than the “functionality of the library: 7.11, library staff must remember that readers.”
facilities, equipment, and so forth (e.g., this single dimension rating as an indi-

266 The Journal of Academic Librarianship


Table 4
Most and Least Important SERVQUAL Dimensions by Library Service Setting
Percent Indicating Dimension is MOST Important1 (%)
Multiple
Study Service Setting N Tangibles Reliability Responsiveness Assurance Empathy Responses
Niteckia ILL 138 1 72 18 3 3 2
Steinb ILL 246 2 79 17 3 0 0
a
Nitecki Reference 92 1 32 37 18 10 2
c
Nitecki Reference 91 1 58 12 24 3 1
a
Nitecki Reserve 109 1 62 21 13 1 2
Nitecki & Hernond Academic library 219 15 62 16 4 5 0
2
Percent Indicating Dimension Is LEAST Important (%)
a
Nitecki ILL 139 83 0 0 3 10 3
b
Stein ILL 243 79 0 0 4 16 0
Niteckia Reference 94 71 4 1 4 19 0
c
Nitecki Reference 92 84 0 0 5 11 0
a
Nitecki Reserve 112 78 3 1 5 10 4
d
Nitecki & Hernon Academic library 219 32 1 4 12 51 0
1
Notes: Based on responses to Question B6.
2
Based on responses to Question B7.
a
Danuta A. Nitecki, An Assessment of the Applicability of SERVQUAL Dimensions as Customer-based Criteria for Evaluating Quality of Services in an Academic Library
(Unpublished doctoral dissertation, University of Maryland, 1995).
b
Joan Stein, “Feedback from a Captive Audience: Reflections on the Results of a SERVQUAL Survey of Interlibrary Loan Services at Carnegie Mellon University Libraries,” in
Proceedings of the 2nd Northumbria International Conference on Performance Measurement in Libraries and Information Services (Newcastle upon Tyne, England: Department
of Information & Library Management, University of Northumbria at Newcastle, 1998), pp. 207–222.
c
Danuta A. Nitecki, “Assessment of Service Quality in Academic Libraries: Focus on the Applicability of the SERVQUAL,” in Proceedings of the 2nd Northumbria International
Conference on Performance Measurement in Libraries and Information Services (Newcastle upon Tyne, England: Department of Information & Library Management, University
of Northumbria at Newcastle, 1998), pp. 181–196.
d
The reference is to the present study.

bust nature of the SERVQUAL score mea- cause they shape the library image of pro- Any significant problems that are uncov-
suring the gap between the ideal viding high quality service, a perception ered should be corrected. Through educa-
expectations and actual service performance that resides with a library’s customers. tion and marketing, staff may also seek to
is a meaningful gauge of quality perfor- If studies do not differentiate partici- manage customer expectations and to
mance for library readers. However, the pants by the service used, some percep- align them with the library’s goals and
foundation of this instrument is customers’ tions might not be based on actual famil- scope of its service programs. The staff
perceptions, which may differ from other iarity with a service. There might be a might also periodically collect spot com-
data collected on service performance. For difference in use of the instrument as a ments on specific service areas and con-
example, through a questionnaire designed gauge of overall service quality and as duct a satisfaction survey, one that centers
for analyzing service quality in terms of one of a more narrowly defined service. around the question reported as C1 in
gaps, readers might record their perception As a result, investigators might, as this Appendix A,28 or post selected questions
about “the availability of computers to use study did, explore overall perceptions and on the library’s Web sites.
the online catalog when needed” as low. expectations, or they might explore opin-
Cost of Replication
Alternative data collection methods might ions about use of a specific service (e.g.,
establish the frequency with which comput- reference or interlibrary loan). The major costs for libraries wanting
ers are not available for use. Staff should Nonetheless, library staff should con- to replicate this survey relate to staff time
review the relationship between both sets of tinue to conduct surveys such as the one in developing the list of expectations to
data.27 If actual measures of performance reported here and seek to reduce the gap survey, in selecting the sample, in prepar-
on the service attributes differ from those of between ideal and actual service attributes ing the surveys for mailing, in logging in
customers’ perceptions of performance, for those expectations they deem impor- the responses, and in entering and verify-
then staff should examine the basis of cus- tant to meet. They might also hold focus ing the data for analysis. (It is assumed
tomers’ perceptions. Hearsay, outdated ex- group interviews with readers, probe the that a library would already possess
periences— ones based on a time period relative importance and unimportance of spreadsheet software and might not need
before service improvements were the attributes grouped through quadrant a statistical analysis software package, if
made— or isolated incidents of dissatisfac- analysis, seek to understand the reasons the motivation for conducting the study is
tion may negatively affect overall impres- for perceived low performance, and so- limited to managing services.) A general
sions of service quality. However, these licit suggestions for how the library can estimate of the amount of labor required
should not be dismissed as unimportant be- improve and surpass quality expectations. to perform each task includes:

July 2000 267


● Deciding on the attributes to study and ploying a survey: mail, in-house, and via CONCLUSION
designing the survey (30 hours); electronic media.30
The survey form and approach presented
● Selecting the study sample, once the The Texas A&M study is developing in this article merit continued examina-
programming to extract the informa- mechanisms by which libraries might attach tion, especially because the findings have
tion was performed (2 hours of report the survey instrument to their Web site and clear implications to library planning and
writing and processing time); encourage customer compliance via e-mail decision making. Individual libraries can
communications. This approach may re- review and modify the service attributes
● Duplicating the questionnaire (the cost quire more effort by the participants than presented in Appendix A and produce an
for the printing of 500 copies was for them to complete a brief print question-
$96.00); Producing copies for the fol- instrument for their local customer input.
naire. Researchers might also experiment Part of the benefit in using the approach
low-up survey to 400 non-respondents
with the use of comment cards containing identified here lays in internal staff re-
increased printing costs ($173.00);
abbreviated information from the question- view and the selection of service at-
● Preparing the surveys for mailing, in- naire reported in Appendix A to focus on tributes for placement on the survey. The
cluding producing and affixing the specific service features. They might ex- staff, it is hoped, become more customer
mailing labels for the sample (40 plore the interconnection between measures or reader focused and more fully grasp
hours); of service quality and of satisfaction. Fur- how each unit of the library contributes to
● Paying for postage, including the cost thermore, staff might replicate the survey on the overall quality of service provided.
of the original mailing (500) and the a regular basis and ascertain whether spe-
return envelopes, (500) plus the sec- cific actions will reduce the breadth of the
ond mailing to 400 non-respondents gap and whether the library can move more
($594, or 1,800 ⫻ $0.33); service attributes into Quadrant 1. Other “Individual libraries can review
● Entering the data from the returned applications of the gap model should also be and modify the service
questionnaires (15 hours); explored; for instance, some investigators attributes presented in
have begun to compare how staff perceive Appendix A and produce an
● Verifying the accuracy of data entry customers will respond in comparison to
(15 hours); and how customers actually respond.31 instrument for their local
● Transcribing and organizing qualita- For those libraries wanting to use a form customer input.”
tive comments (20 hours). on an on-going basis, researchers should
experiment with the use of a 10-point scale
By having access to the survey re- so that libraries can more precisely monitor Whatever decisions that a library makes
ported in Appendix A, library staff should the impact of service improvements on ser- about which service attributes to probe, the
save time in selecting candidate service vice quality. And, finally, further work survey form should not be longer than the
attributes to investigate and in designing a ought to address the application of the five one used in this study. The goal is to probe
questionnaire. They can definitely adapt the service attributes of greatest potential
SERVQUAL dimensions in libraries, in-
the form to meet local needs (see Note 22 value to both actual customers and service
cluding the reorganization and expansion of
for examples of other service attributes). providers, and not to overwhelm respon-
Moreover, if the staff prefer to conduct the dimensions. At present, they inconsis-
tently and, thus, inadequately represent ac- dents with choices, thereby raising false
in-house surveys, they save on mailing hopes or expectations and producing an in-
costs but increase staff time for the distri- ademic libraries as service organizations.
The next phase of developing an assess- strument where the scores may lack reliabil-
bution and collection of survey forms. ity. Although, over time, the set of selected
ment method is to explore how library staff
use the data obtained through the applica- attributes may change, they should continue
FURTHER RESEARCH to reflect areas in which the library is com-
tion of the survey instrument. How can in-
In a presentation of a pilot project using formation be organized and presented to the mitted to make service improvements as
SERVQUAL in 12 member institutions of necessary to respond to its customers’ ex-
staff to motivate them in their efforts to
the Association of Research Libraries pectations.
improve service? What preparation is
(ARL), Bruce Thompson, professor of ed- The importance of conducting an assess-
needed to ensure that staff will embrace
ucational psychology and distinguished re- ment such as the one described in this arti-
search scholar, Texas A&M University, de- assessment data as part of the accountability cle is not limited to calculating measures of
clared that he and his colleagues were movement and not react defensively to the which service attributes are important to
striving for a 20 –30% return rate, using a findings? Is assessment that is conducted readers or which ones readers perceive as
Web-based questionnaire.29 The study re- centrally more or less meaningful or rele- delivered well. We also doubt the impor-
ported here had a higher return rate but, vant to the staff than performance moni- tance of such an investigation lies in any
although still representative of the sample, tored by individual service units? What implied comparison or ranking of service
was lower than what we would have pre- amount of detail is relevant to the service quality across libraries. The importance, we
ferred. Clearly, there is need for research staff? What types of comparison over time believe, rests with what will be done with
that explores how to improve response rates and across service units are valuable? How the insights gained. The quantified mea-
in survey research related to service quality. does feedback on gap perceptions influence sures of expectations and service quality are
Also, there is need for more case studies staff ambitions for surpassing the meeting only valuable if staff discuss them and man-
that continue to review and refine the ap- of customer expectations? Is there an effi- agers help to establish action plans and
proach presented in this article as well as to cient way to collate written remarks on monitor service improvements. Open-ended
experiment with different methods of de- questionnaires? questions and commentaries in the margin

268 The Journal of Academic Librarianship


of the instrument provide opportunities for 5. Ibid. sure of Service Quality in Academic Li-
respondents to share potential problems and 6. Zeithaml, Parasuraman, & Berry, Deliver- braries,” p. 185.
compliments. The compliments should be ing Quality Service. See also Danuta A. 18. Nitecki, “Assessment of Service Quality
passed along to the staff and the problems Nitecki, “Changing the Concept and Mea- in Academic Libraries,” p. 184.
sure of Service Quality in Academic Librar- 19. Hernon & Whitman, Delivering Satisfac-
should be reviewed and, where possible,
ies,” Journal of Academic Librarianship 22 tion and Service Quality; Hernon & Alt-
resolved. (May 1996): 182. man, Service Quality in Academic Librar-
The survey instrument, thus, becomes an 7. Ibid. ies; Hernon & Altman, Assessing Service
important communications channel be- 8. See Peter Hernon & Ellen Altman, Service Quality; Hernon & Calvert, “Methods for
tween customers and staff. There is a need Quality in Academic Libraries (Norwood, Measuring Service”; Calvert & Hernon,
to create a method of communications back NJ: Ablex, 1996); Peter Hernon & Ellen “Surveying Service Quality within Univer-
to the customers through library education Altman, Assessing Service Quality: Satis- sity Libraries.”
and marketing channels. The Gaps Model fying the Expectations of Library Custom- 20. Nitecki, “Changing the Concept and Mea-
of Service Quality contends that there are ers (Chicago, IL: American Library Asso- sure of Service Quality in Academic Li-
several contributors to an institution’s abil- ciation, 1998); Peter Hernon & Philip J.
braries,” p. 188.
ity to deliver quality service. We have only Calvert, “Methods for Measuring Service
21. Cook & Thompson, “Reliability and Valid-
Quality in University Libraries in New
explored here the one related to understand- Zealand,” Journal of Academic Librarian-
ity of SERVQUAL Scores Used to Evaluate
ing the customer’s expectations and percep- ship 22 (September 1996): 387–391; Perceptions of Library Service Quality.”
tions of service quality. Of critical impor- Philip J. Calvert & Peter Hernon, “Sur- 22. Hernon & Altman, Assessing Service
tance is the accurate communication of the veying Service Quality within University Quality, pp. 108 –112.
results of this understanding in the form of Libraries,” Journal of Academic Librari- 23. James Lynch, Robert Carver, Jr., & John
service commitments to customers. A li- anship 23 (September 1997): 408 – 415. Michael Virgo, “Quadrant Analysis as a
brary needs to explain why it might not be 9. Zeithaml, Parasuraman, & Berry, Deliver- Strategic Planning Technique in Curricu-
able to resolve all problems identified in ing Quality Service, p. 26. lum Development and Program Market-
10. Nitecki, “Changing the Concept and Mea- ing,” Journal of Marketing for Higher Ed-
surveys. Clearly, complaints represent op-
sure of Service Quality in Academic Li- ucation 7 (1996): 17–32.
portunities to improve service and to inter-
braries,” pp. 184 –185. 24. Thomas A. Childers & Nancy A. Van
act directly with dissatisfied customers. House, What’s Good? Describing Your
11. Danuta A. Nitecki, “Assessment of Service
Complaints are valued feedback within a Quality in Academic Libraries: Focus on the Public Library’s Effectiveness (Chicago, IL:
library culture of service quality that exists Applicability of the SERVQUAL,” in Pro- American Library Association, 1993), p. 26.
when staff divorce customer criticism from ceedings of the 2nd Northumbria Interna- 25. See Hernon & Altman, Assessing Service
personal judgments of performance and, in- tional Conference on Performance Mea- Quality; Hernon & Altman, Service Qual-
stead, focus on continually redesigning ser- surement in Libraries and Information ity in Academic Libraries. See also Nicola
vices to exceed customer expectations of Services (Newcastle upon Tyne, England: Harwood & Jillene Bydder, “Student Ex-
what the library deems is its program. Department of Information & Library Man- pectations of, and Satisfaction with, the
A culture of service quality assessment agement, University of Northumbria at University Library,” Journal of Academic
provides opportunities to demonstrate to Newcastle, 1998), p. 185. The Maryland Librarianship 24 (March 1998): 161–171.
customers how what the staff learns about study mentioned in the quotation refers to 26. See Hernon & Whitman, Delivering Sat-
Danuta A. Nitecki, An Assessment of the isfaction and Service Quality. Clearly, the
customers’ expectations and perceptions
Applicability of SERVQUAL Dimensions as study reported in this article deals with
helps to shape the service that libraries Customer-based Criteria for Evaluating
provide and the commitments that librar- service quality, and the complete interre-
Quality of Services in an Academic Library. lationship between service quality and sat-
ians make to their customers. Such oppor- Unpublished doctoral dissertation, Univer- isfaction were not explored here.
tunities should not be ignored. sity of Maryland, 1995.
27. See Hernon & Altman, Assessing Service
12. Syed Saad Andaleeb & Patience L. Sim-
NOTES AND REFERENCES Quality, Chapter 9.
monds, “Explaining User Satisfaction
28. See Hernon & Whitman, Delivering Sat-
1. Valarie A. Zeithaml, A. Parasuraman, & with Academic Libraries: Strategic Impli-
cations,” College & Research Libraries isfaction and Service Quality.
Leonard L. Berry, Delivering Quality Ser-
59 (March 1998): 156 –167. 29. Presentation before an invited audience at
vice: Balancing Customer Perceptions
and Expectations (New York: The Free 13. Emin Babakus & Gregory W. Boller, “An the American Library Association Midwin-
Press, 1990), p. 16. Empirical Assessment of the SERV- ter Conference, San Antonio, January 15,
2. Ibid. QUAL Scale,” Journal of Business Re- 2000.
3. Carol A. Reeves & David A. Bednar, “De- search 24 (May 1992): 253–268. 30. For guidance in the conduct of such a
fining Quality: Alternatives and Implica- 14. Colleen Cook & Bruce Thompson, “Reli- survey, see Hernon & Altman, Assessing
tions,” Academy of Management Review ability and Validity of SERVQUAL Service Quality; Hernon & Whitman, De-
19 (July 1994): 437. Scores Used to Evaluate Perceptions of livering Satisfaction and Service Quality.
4. Peter Hernon & John R. Whitman, Deliv- Library Service Quality,” Journal of Aca- 31. See Susan Edwards & Mairéad Browne,
ering Satisfaction and Service Quality: A demic Librarianship 26 (July 2000): 248. “Quality in Information Services: Do Users
Customer-Based Approach for Libraries 15. Ibid. and Librarians Differ in Their Expectations,
(Chicago, IL: American Library Associa- 16. Ibid. Library & Information Science Research 17
tion, 2000). 17. Nitecki, “Changing the Concept and Mea- (Spring 1995): 163–182.

July 2000 269


270 The Journal of Academic Librarianship
July 2000 271
272 The Journal of Academic Librarianship
Appendix B that library services at Yale possess new Quadrant 2, reflecting important
Use of Quadrant Analysis them all. To differentiate the relative attributes for readers that the library
In this study, the mean scores for importance of the service attributes to does not deliver above the 5.5 level of
responses to the ranking of ideal expec- prioritize efforts to improve services, a performance. Among the attributes
tations and of Yale’s library perfor- higher threshold of importance was set. deemed by readers as less important (be-
mance on each of the 40 statements of For the purposes of analysis, the quad- low the 5.5 level), only three were not
service attributes were plotted for the rant axes were relocated to intersect at perceived as part of the library’s service
quadrant analysis. If the intersection of the 5.5 value for each dimension. This strengths; these fell within Quadrant 4.
the two axes is left at the neutral mid- placed 11 attributes in the newly de- No attributes fell within Quadrant 3;
point value of 3.5 for each dimension, fined Quadrant 1; these correspond readers did not see any service factors as
then all attributes in this study fall with an importance to readers of 5.5 or unimportant but performed relatively well
within Quadrant 1. This suggests initial more on a seven-point scale and with a by the library.
impressions that readers value all the perception that they exist in the library’ The service attributes are plotted in
service factors as important for excel- services ranked along equally high lev- Figure 2, and Table 1 lists those attributes
lent service quality and that they find els. Twenty-six attributes fell into the falling in Quadrants 1, 2, and 4.

July 2000 273

You might also like