Вы находитесь на странице: 1из 10

Satisfaction Formation Processes in Library Users:

Understanding Multisource Effects


by Xi Shi, Patricia J. Holahan, and M. Peter Jurkat

This study explores whether nderstanding satisfaction forma- ture. One example of such products is
disconfirmation theory can
explain satisfaction formation
processes in library users. Both
U tion processes in library users has
been a concern of researchers and
practitioners alike in the field of library
restaurants.2 The product of a restaurant
is its food. However, the customer satis-
faction judgment is based not only on food
and information sciences. Although user quality (the product itself), but also factors
library users’ needs and satisfaction is not an unfamiliar topic in such as speed of service, friendliness of
library science, there is not yet a common- waiters or waitresses, and so forth. Like-
expectations are investigated as ly accepted theoretical model that predicts wise, within library science research, in-
disconfirmation standards. library user satisfaction. Well-developed formation can be seen as a consumable
Overall library user satisfaction and tested theories of satisfaction are product that can only be consumed togeth-
is predicted to be a function of needed to guide the practical aspects of er with certain information delivery sys-
library management. The current study tems and/or services.
two independent sources— applies a theoretical model of customer Thus, we propose that satisfaction for
satisfaction with the satisfaction/dissatisfaction (CS/D) from library users is a function of multiple
information product received the field of marketing to the study of sources—the customer’s satisfaction with
and satisfaction with the satisfaction formation processes in library the information product(s) received as
users. This study is a solid step in linking well as satisfaction with the information
information system and library customer satisfaction/dissatisfaction re- system and library services utilized to
services used to retrieve the search to the fields of library and informa- obtain the information product. This prop-
information product. Both tion sciences and provides a theoretical osition is consistent with the propositions
framework for understanding satisfaction made by other library science researchers
sources are hypothesized to formation processes in library users as and the findings from several satisfaction
contribute independently to consumers of information services and studies. For example, Singh3 studied sat-
satisfaction in library users. products. isfaction in the health care industry and
found that overall patron satisfaction was a
CONSUMER SATISFACTION/ result of three sources—satisfaction with
DISSATISFACTION THEORY the physician, satisfaction with the hospi-
tal, and satisfaction with the insurance
Library User Satisfaction and provider. Halstead and Hartman4 studied
Multisource Effects alumni satisfaction and found overall sat-
In this study, information is investigated isfaction was a result of satisfaction with
as a consumable product. Library users are two independent sources—namely, satis-
customers who are having a purchasing faction with intellectual environment and
experience. Accordingly, information sear- satisfaction with employment preparation.
ching and retrieval activities are viewed as Similarly, library science researchers have
‘‘purchasing’’ experiences of library users. suggested that satisfaction with services
Xi Shi, As with many consumables, information (such as ‘‘user friendliness’’ of the infor-
SUNY Rockland Community as a product is defined by two components: mation search system, speed of retrieval,
College, Suffern, NY 10901, USA (1) the information system/services1 that etc.) and satisfaction with the information
<xshi@sunyrockland.edu>; the customer utilizes to access and retrieve obtained (such as its content and quality)
Patricia J. Holahan, the information product and (2) the infor- might be two distinct entities.5 Consistent
Stevens Institute of Technology, mation product itself. with this proposition are the research find-
Hoboken, NJ 07030, USA; Research on products bundled together ings by Applegate.6 Applegate found that
M. Peter Jurkat, with certain services in a way similar to sometimes the library users would claim to
Stevens Institute of Technology, information products is found in the cus- be ‘‘satisfied’’ even when their questions
Hoboken, NJ 07030, USA. tomer satisfaction/dissatisfaction litera- were not answered. Applegate called this

122 The Journal of Academic Librarianship, Volume 30, Number 2, pages 122–131
phenomenon ‘‘emotional satisfaction’’ as subscription databases, which enables the positive disconfirmation, satisfaction
opposed to ‘‘material satisfaction.’’ user to navigate the databases comfortably occurs. When performance is less than
It is noteworthy that research on deter- and conduct a thorough search easily. In expectations (or other prepurchase stan-
minants of satisfaction within the library the second case, the library user may dard), resulting in negative disconfirma-
science literature has most often not dis- experience a given level of satisfaction tion, dissatisfaction occurs. Confirmation
tinguished between satisfaction with the because the information they received occurs when performance and expecta-
information product and satisfaction with was current and relevant to their problem. tions match, resulting in moderate satis-
the information system/services. That is, In the first case, satisfaction is a result of faction or indifference. Thus, the extent to
researchers have assessed only the user’s user’s experience with the information which a customer experiences satisfaction
satisfaction with the information prod- system/services. In the second case, satis- or dissatisfaction is related to the size and
uct(s) received, or they have assessed faction results from the user’s perceptions direction of disconfirmation.11 We pro-
only the quality of information system/ of the information product itself. Each of pose that disconfirmation can explain sat-
services that users accessed when seeking these sources of satisfaction is believed to isfaction formation processes in library
information.7 Or, they have chosen to be independent, that is, one does not users. We now turn to a discussion of the
operationalize satisfaction as a global depend on the other. Accordingly, we offer components of the model-perceived per-
construct, confounding users’ satisfaction the following hypotheses. formance and disconfirmation standards
with the information product and infor- (see Fig. 1).
mation system/services.8 In the first two Hypothesis 1: Increases in satisfaction
cases, we contend that the satisfaction with the information product are related to Performance
construct is underspecified. In the third increases in the user’s overall satisfaction. In disconfirmation theory, performance
case, we contend the satisfaction con- is defined as the customer’s subjective
struct is confounded. Assessing satisfac- Hypothesis 2: Increases in satisfaction
perception of the quality of the product
tion globally clouds our understanding of with the information system/services are (or service) after it is consumed. Perfor-
the satisfaction construct and its determi- related to increases in the user’s overall mance, moreover, is seen as multifaceted.
nants. Moreover, it is possible that the satisfaction. According to the marketing literature, cus-
tomer satisfaction is not determined by
Satisfaction Formation evaluation of any one factor or component
‘‘Assessing satisfaction Disconfirmation theory9 is the most of performance.12 Instead, customer satis-
popular and widely used model for pre- faction is believed to be the result of
globally clouds our dicting customer satisfaction/dissatisfac- customer evaluations of several compo-
understanding of the tion. The basic concepts that compose nents or attributes of performance.
satisfaction construct and disconfirmation theory are shown in Fig. In the field of library and information
its determinants.’’ 1. According to disconfirmation theory, science, attributes used to assess the ‘‘per-
before shopping, customers have some formance’’ of information products in-
prepurchase standard(s) in their minds clude the perceived accuracy, complete-
(such as performance expectations) that ness, currency, precision, relevance, and
determinants of satisfaction with the in- guide their purchasing activities. After reliability of the information product.13
formation system/services differ from the purchasing a product (service), customers Accordingly, library users are believed to
determinants of satisfaction with the in- evaluate the performance of the product evaluate these attributes of the information
formation product. Unless our research (service) against this prepurchase stan- received (e.g., its’ accuracy, completeness,
recognizes and conceptualizes the multiple dard. Accordingly, CS/D is defined as precision, relevance) against some precon-
sources of satisfaction separately and the consumer’s response to the evaluation sumption or disconfirmation standard.
comprehensively, we may not be able to of a perceived discrepancy between the When the user’s judgment of the product
develop adequate models of satisfaction prepurchase performance standard and the performance (i.e., the product attributes)
formation. actual performance of the product (ser- exceeds their prepurchase standard(s), sat-
Thus, this research proposes that satis- vice) after its consumption.10 When per- isfaction results. Conversely, when the
faction in library users is a function of two formance is greater than expectations (or user’s evaluation of the product perfor-
sources—satisfaction with the information other prepurchase standard), resulting in mance (i.e., product attributes) falls short
product and satisfaction with the informa-
tion system/services that provide the in-
formation product. We propose that each Figure 1
of these sources of satisfaction contributes Conceptual Model of CS/D.
independently to the library user’s overall
satisfaction. Consider two different cases.
In the first case, a library user may expe-
rience a given level of satisfaction because
they found the information system/serv-
ices used to retrieve the information prod-
uct convenient to access, easy to use, etc.
For example, a library may have a well-
developed home page that provides clear
explanations of how to use the library

March 2004 123


of their prepurchase standard(s), dissatis- ogy available in this small college may the library user’s expectations, the greater
faction results. not be as sophisticated as what he is used the library user’s satisfaction with the
In a similar fashion, several attributes to in his home institution. Therefore, he information product.
have been identified as comprising satis- may expect more limited access to infor-
faction with information system/services. mation and supporting materials com- Hypothesis 5: The more the information
Attributes that have been studied as com- pared to what he could retrieve from his system/services performance is perceived
ponents of satisfaction with information home library. We propose that both li- to fulfill the library user’s needs, the
system/services include response time, at- brary users’ needs and expectations func- greater the library user’s satisfaction with
titude of service staff (e.g., librarians), the tion as disconfirmation standards and the information system/services.
user friendliness of the information sys- both play a role in explanation of satis-
tem, and convenience of access. faction formation processes in informa- Hypothesis 6: The more the information
tion users. system/services performance is perceived
Disconfirmation Standards In summary, we propose that the cur- to fulfill the library user’s expectations,
rent models of satisfaction are underspe- the greater the library user’s satisfaction
Within disconfirmation theory, discon- cified since needs have not been included with the information system/services.
firmation standards are the customer’s in the model construction. We predict Fig. 2 presents the proposed theoret-
prepurchase or preconsumption compari- ical model. In summary, we propose that
son standards.14 Although the dominant overall library user satisfaction is influ-
model of CS/D in the field of marketing enced by the user’s satisfaction with the
continues to be the model based on the ‘‘...we propose that the
information product itself and with the
‘‘disconfirmation of expectations,’’ dis- current models of information system/services that provide
confirmation standards other than perfor- satisfaction are the information product. Product satis-
mance expectations have been investigat- underspecified since needs faction and information system/services
ed. The following are some examples. satisfaction are, in turn, a function of
Barbeau15 reported a study of student have not been included in
the degree to which the information
satisfaction with a marketing course. In the model construction.’’ product and the information system/serv-
his study, disconfirmation of past experi- ices meet or exceed customer needs and
ence, disconfirmation of expectations, expectations.
and disconfirmation of needs were tested satisfaction is highest when performance
as three separate disconfirmation stand- exceeds both the user’s expectations and RESEARCH METHODS
ards. Spreng and Olshavsky16 compared needs. Given the foregoing discussion, we
customer needs and customer expecta- Research Sample
offer the following hypotheses:
tions as disconfirmation standards and The research sample consists of 105
found customer needs to be a powerful Hypothesis 3: The more the information faculty and administrators from eight
predictor of satisfaction. Their study was product performance is perceived to fulfill accredited science and engineering col-
replicated by Spreng et al.17 who also the library user’s needs, the greater the leges and universities located in the north-
found that disconfirmation of consumer library user’s satisfaction with the infor- east United States. The eight colleges and
needs explained consumer satisfaction mation product. universities were selected primarily as a
better than disconfirmation of consumer convenience sample. Five hundred and
expectations. Hypothesis 4: The more the information forty-five questionnaires were sent out.
In library science, customer needs as product performance is perceived to fulfill One hundred and twenty-one usable ques-
well as customer expectations are often
assumed to affect satisfaction.18 In the
present study, library users’ needs and Figure 2
expectations are investigated as disconfir- Hypothesized Model.
mation standards. Library users’ ‘‘needs’’
refer to what the user wants or desires to
receive from their information search. The
term ‘‘expectations’’ as a disconfirmation
standard refers to what the library users
believe they will receive from their infor-
mation search. To illustrate the difference
between needs and expectations, the fol-
lowing example is given. A professor
working in an academic institution with
extensive library collections and sophis-
ticated computer search technology is
invited to teach as a visiting scholar by
a small college. This professor needs
supporting materials for his teaching
and research at his new position. How-
ever, the library collections and technol-

124 The Journal of Academic Librarianship


tionnaires were returned, for a response sion—the degree of exactness in the infor- Spreng et al.22 Using a five-point scale,
rate of 22 percent. This response rate is mation; (3) relevance—the degree of we asked respondents to indicate how
similar to or better than that reported in pertinence or congruence of the informa- important (1 = not important to 5 = very
other studies of satisfaction formation tion relative to the interests of the user; (4) important) each of the five information
processes.19 detail—the amount and depth of the product performance attributes was to
knowledge that is delivered to the user them in the information for which they
Data Collection Methods by the information; and (5) appropriate- were searching. The five product perfor-
Questionnaires were sent directly to ness—the extent to which the format, mance attributes included the accuracy of
faculty of the selected colleges and uni- language, and comprehension levels of the information, the appropriateness of
versities and to the head librarian at the the information are suitable for the user. the information, the level of detail in
eight colleges and universities. The faculty The five information system/services the information, the precision of the
wh o re ceiv ed que stionn aires w as performance attributes included the fol- information, and the relevance of the
requested to complete the questionnaire lowing: (1) ease of searching—the techni- information.
for an upcoming information search. The cal aspects of information storage and
librarians were requested to hand out the retrieval systems; (2) user-friendliness— User Expectations (Information
questionnaires to faculty and administra- the mechanical aspects of the information Product)
tors who were initiating an information systems including the interface, display, Our measure of user expectations for
search. A sheet of detailed instructions format, and navigational design of the the information product, adapted from
accompanied all questionnaires. The ques- system; (3) timeliness—the time required Spreng et al.,23 asked respondents to eval-
tionnaire contained two parts. Respond- to obtain the needed information from the uate the extent to which they expected
ents were instructed to complete ‘‘Part I: point where the information search was each of five performance attributes—ac-
Presearch Questions’’ before they engaged initiated to the point where the information curacy, appropriateness, detail, precision,
in searching activities, that is, precon- is received by the user; (4) staff or knowl- and relevance—to be present in the infor-
sumption. Respondents were instructed edge skills—the capability and expertise mation for which they were searching.
to complete ‘‘Part II: Postsearch Ques- of the library staff in assisting the infor-
tions’’ after the search was complete, that mation seeker in finding the needed infor- User Needs (Information System or
is, postconsumption. A preaddressed, mation; and (5) helpful staff attitude—the Services)
stamped envelope was provided for the library staff’s willingness to assist the Our measure of user needs, also adap-
return of the questionnaire. information seeker in finding the needed ted from the research by Spreng et al.,24
Part I of the questionnaire asked res- information. asked respondents to indicate how impor-
pondents to record their needs and tant (1 = not important to 5 = very impor-
expectations for an upcoming information User Needs (Information Product) tant) each of the five performance attri-
search. Needs and expectations for the Our measure of user needs for the butes was to them in the information
information product as well as for the information product was adapted from system/services they would access during
information system/services were soli-
cited. In Part II, respondents rated the
performance of the information product Table 1
received and the information system/serv- Descriptive Statistics for the Measured Variables
ices accessed and responded to the scales
used to measure disconfirmation (see Mean SD Min Max No. of a
Items
below).
Overall customer 3.91 0.94 1.00 5.00 1 N/A
Measures satisfaction
Five information product performance Satisfaction with product 3.81 0.96 1.00 5.00 5 .96
attributes, derived from the User Infor-
mation Satisfaction (UIS) survey devel- Satisfaction with service 3.73 0.93 1.00 5.00 5 .91
oped by Ives et al.20 and revised by User needs (product) 4.46 0.43 3.20 5.00 5 .65
Baroudi and Orlikowski,21 were used to User expectations (product) 3.27 0.50 2.20 4.00 5 .84
assess needs, expectations, needs discon-
firmation, expectations disconfirmation, Need disconfirmation 3.77 0.76 1.00 5.00 5 .88
and satisfaction with the information (product)
product. Five information system/services Expectation disconfirmation 3.80 0.76 1.00 5.00 5 .87
performance attributes also derived from (product)
the revised UIS were used to assess User needs (system or service) 4.36 0.54 2.60 5.00 5 .68
needs, expectations, needs disconfirma-
tion, expectations disconfirmation, and User expectations 4.00 0.76 2.00 5.00 5 .85
satisfaction with the information system/ (system of service)
services. Need disconfirmation 3.96 0.70 1.00 5.00 5 .84
The five information product perfor- (system or service)
mance attributes included the following:
Expectation disconfirmation 3.97 0.72 1.00 5.00 5 .86
(1) accuracy—the extent to which the
(system or service)
information is correct and true; (2) preci-

March 2004 125


their information search. The five service five information product (information items adapted from the research by Patter-
performance attributes included the fol- system/services) attribute ratings were son et al. On a five-point scale, res-
lowing: (1) ease of searching, (2) user- then averaged, which yielded an expect- pondents were asked the extent to which
friendliness of search tools, (3) timeliness ations disconfirmation score for the in- they agreed or disagreed (1 = strongly
with which they obtained they informa- formation product and an expectations disagree to 5 = strongly agree) with each
tion, (4) staff knowledge and skills, and disconfirmation score for information of five statements. Sample statements in-
(5) helpfulness of the staff. system/services. cluded, ‘‘I can find what I am looking for
because of the available information serv-
User Expectations (Information Satisfaction with Information Product ices’’, and ‘‘The information services here
Services) Information product satisfaction was helped me find what I was looking for.’’
Our measure of user expectations measured using five items adapted from
asked respondents to evaluate the extent the research by Patterson et al.27 On a five- Overall Library User Satisfaction
to which they expected each of five per- point scale, respondents were asked the Overall satisfaction is defined as the
formance attributes—ease of searching, extent to which they agreed or disagreed library user’s subjective summary evalua-
user-friendliness, timeliness, staff knowl- (1 = strongly disagree to 5 = strongly tion of their entire information search acti-
edge and skills, and helpful staff atti- agree) with each of five statements. Sam- vity as one information ‘‘purchase’’ expe-
tude—to be present in the information ple statements include, ‘‘I am very satis- rience. The evaluation includes the user’s
system/services they would access during fied with the information I found from this experience with the information product
their information search. This measure search’’, and ‘‘The information I found is and with using available information sys-
was also adapted from the research by just right to solve my problems.’’ tems and services in retrieving the infor-
Spreng et al.25 mation. Overall satisfaction was ope-
Satisfaction with Information System rationalized by a single item measured
Operationalization of Needs or Services on a five-point scale. The item asked
Disconfirmation and Expectations Satisfaction with the information sys- respondents ‘‘Overall, how satisfied were
Disconfirmation tem/services was measured using five you with your information search?’’ where
We measured disconfirmation using
the procedure recommended by Spreng
et al.26 Needs disconfirmation and expect- Table 2
ations disconfirmation for the information Correlations Among the Measured Variables
product and for the information system/
services were measured in a similar way. Variables 1 2 3 4 5 6 7 8 9 10 11
Needs disconfirmation was measured for 1. User needs 1.00
each performance attribute by asking, (product)
‘‘For each of the five attributes listed,
please tell us if the information you found 2. User expectations .26* 1.00
(or the information system/services you (product)
utilized) was ‘5 = exactly what I needed,’ 3. Need disconfirmation .22* .08 1.00
‘3 = somewhat different from what I (product)
needed,’ ‘1 = extremely different from 4. Expectation .19 .04 .72** 1.00
what I needed.’’’ Thus, this question disconfirmation
asked the respondents for their subjective (product)
assessment of the discrepancy between
their needs and the perceived perfor- 5. User needs .36* .30* .20* .16 1.00
mance. The five information product (in- (system or service)
formation system/services) attribute 6. User expectations .06 .50* .07 .07 .55* 1.00
ratings were then averaged, which yielded (system or service)
a needs disconfirmation score for the
7. Need disconfirmation .20* .18 .52* .45** .09 .07 1.00
information product and a needs discon-
(system or service)
firmation score for information system/
services. 8. Expectation .10 .20* .33* .51* .11 .01 .72** 1.00
Expectation disconfirmation was oper- disconfirmation
ationalized in the same way. Respondents (system or service)
were asked for their subjective assess- 9. User satisfaction .20* .09 .72** .69** .18 .10 .53* .44* 1.00
ment of the discrepancy between what with product
they expected and what they received.
That is, respondents were asked, ‘‘Please 10. User satisfaction .21* .05 .58* .57* .08 .08 .67** .55* .70** 1.00
tell us if the information you found (or with system or
the information system/services you uti- service
lized) was ‘5 = exactly what I ex- 11. User satisfaction .19 .01 .66* .55* .18 .11 .58** .46* .84** .76** 1.00
pected,’ ‘3 = somewhat different from overall
what I expected,’ ‘1 = extremely differ- Notes: *P < .05.
ent from what I expected.’’’ Again, the **P < .01.

126 The Journal of Academic Librarianship


Table 3 unique contribution to explained variance
Stepwise Regression: Overall Library User Satisfaction Regressed on in overall customer satisfaction (see Table
Satisfaction with Information Product and Satisfaction with Information 3). Service satisfaction was entered as a
System/Services predictor in the first round but was exclud-
ed in the second round. Product satisfac-
Predictors B SE b t Variables tion accounted for 77 percent of the
removed
variance in the dependent variable ( F =
Product satisfaction .64*** 0.07 .65 9.37 160, P < .001). This finding suggests that
System or service .30 0.07 .29 4.25 X satisfaction with the information product
satisfaction may be a better predictor of overall library
user satisfaction than satisfaction with the
Notes: F = 160.30***; R2 = .77; adjusted R2 = .77.
***P < .001.
information system/services accessed to
retrieve the information product.

1 = not satisfied at all and 5 = very relative contribution of satisfaction with


satisfied. the information product and satisfaction ‘‘...satisfaction with the
with services in determining overall sat- information product may
RESEARCH RESULTS isfaction, stepwise regression analysis be a better predictor of
will be used.
Descriptive Statistics and Correlation Measures of need disconfirmation and overall library user
The mean, range, standard deviation, expectations disconfirmation are highly satisfaction than
number of items per scale, and, where correlated (r = .72) for both products and satisfaction with the
appropriate, coefficient alpha for the mea- system or services. Again, stepwise re-
gression analysis will be used in tests of
information system/
sured variables are presented in Table 1.
All variables were measured using Likert the hypotheses where need disconfirma- services accessed...’’
scales that ranged from 1 to 5, with 5 tion and expectation disconfirmation are
indicating a greater degree of the measured included as predictors. The use of stepwise
variable. regression analysis should allow us to Tests of Hypotheses 3 and 4
Coefficient alphas for the measured separate out the unique contribution each Hypotheses 3 and 4 predicted that both
variables ranged from .65 to .96. Coeffi- predictor makes to explained variance in needs and expectations disconfirmation
cient alpha is a measure of reliability or the dependent measure. predicted of satisfaction with the informa-
internal consistency and indicates whether tion product. To test Hypotheses 3 and 4,
Tests of Hypotheses stepwise regression was conducted. Satis-
respondents answer the items that com-
pose a scale in a consistent fashion. Co- Test of Hypotheses 1 and 2 faction with the information product was
efficient alphas for the scales assessing Hypotheses 1 and 2 proposed that o- regressed on needs disconfirmation and
user needs for the information product verall user satisfaction of information expectations disconfirmation for the infor-
and user needs for information services products or services is a result of the mation product (see Table 4). User needs
were .65 and .68, respectively. According independent effects of satisfaction with and user expectations were included in the
to Cohen and Cohen,28 these reliabilities the information product and satisfaction regression equation as control variables—
are acceptable when doing research in the with information system/services accessed that is, it was necessary to demonstrate
behavioral and social sciences. The stan- to obtain the information product. Given that it was the disconfirmation of user
dard deviations for the scales assessing the high correlation between satisfaction needs and expectations rather than user
user expectations for the information with product and satisfaction with serv- needs or expectations alone that predicted
product, user needs for the information ices, a stepwise regression was run to test satisfaction. As shown in Table 4, user
product, and user needs for information whether each of these predictors made a needs and expectations were removed
system/services indicate that there may be
some restriction of range on these varia-
bles. This restriction in range may make it Table 4
more difficult to detect a difference when Stepwise Regression: Satisfaction with Information Product Regressed
indeed one may exist. on Needs and Expectations Disconfirmation Controlling for User Needs
Intercorrelations among the measured and Expectations
variables are shown in Table 2. Satisfac-
tion with the information product and Predictors entered B SE b t Variables
removed
satisfaction with information services are
correlated .70. This result suggests that User needs (product) X
library users, who feel satisfied with the User expectations (product) X
information product, feel satisfied with
information services and vice versa. Need disconfirmation (product) .55*** 0.12 .42 4.43
Thus, these two variables may not be Expectation disconfirmation (product) .51*** 0.12 .40 4.22
independent predictors of overall satisfac- Notes: F = 67.11***; R2 = .58; adjusted R2 = .57.
tion as proposed. To separate out the ***P < .001.

March 2004 127


Table 5 hard to navigate, etc.). The problem with
Stepwise Regression: Satisfaction with Information System/Services this prescription is that the users may
Regressed on Needs and Expectations Disconfirmation Controlling for never get to experience the quality of
User Needs and Expectations the information product if they cannot
navigate the delivery system. Thus, it
Predictors B SE b t Variables would appear that library information
removed
management should consider developing
User needs X both the quality of information products
(information system/service) through collection development as well as
User expectations X improvement of the information delivery
(information system/service) systems or services.
Second, this study explored the appli-
Needs disconfirmation .87*** 0.11 .67 8.11 cability of the disconfirmation paradigm in
(information system/service) explaining satisfaction formation process-
Expectations disconfirmation X es in library users. Our results support that
(information system/service) it is the disconfirmation of user needs and
Note: F = 65.72***; R2 = .46; adjusted R2 = .45.
expectations and not user needs or expect-
***P < .001. ations alone that determine library user
satisfaction. That is, satisfaction or dissat-

from the model in Step 2. However, both formation process for library users.
needs disconfirmation (b = .42, P < .001) Within the disconfirmation paradigm,
and expectations disconfirmation (b = .40, both library user’s needs and expect-
‘‘...disconfirmation of user
P < .001) were retained. These predictors ations were investigated as disconfirma- needs and expectations and
jointly accounted for 57 percent of the tion standards. Overall, library user not user needs or
variance in the dependent variable ( F = satisfaction was conceptualized as com- expectations alone that
67.11, P < .001). These findings support posed of two independent sources—sat-
Hypotheses 3 and 4—that is, both needs isfaction with the information product determine library user
and expectations function as disconfirma- received and satisfaction with the infor- satisfaction.’’
tion standards and predict satisfaction with mation system and services used to
the information product. retrieve the information product. Both
sources were hypothesized to contribute
Tests of Hypotheses 5 and 6 independently to overall library user isfaction results only when consumers
Hypotheses 5 and 6 predicted that needs satisfaction. compare the performance of what they
disconfirmation and expectations discon- Our results indicate that satisfaction received against some ‘‘prepurchase’’
firmation predicted of satisfaction with the with the information product may be or disconfirmation standard. This result
information system/services. Stepwise re- more important for overall satisfaction is consistent with earlier researches
gression was conducted to test whether when compared to satisfaction with the and supports the notion that needs
needs disconfirmation and expectations information system/service. This finding and expectations affect satisfaction
disconfirmation made a unique contribu- is consistent with Pitt et al.29 who contend only indirectly via the disconfirmation
tion to explained variance in user satisfac- that information is the dominant concern process.30
tion with information system/services (see of the user. Thus, it may be that finding Third, our research explored discon-
Table 5). User needs and user expectations the information sought is more important firmation standards other than expecta-
were included in the regression equation as to the library user than the ease with tions as predictors of satisfaction in
control variables. As can be seen from which the information is obtained. In library users. Our results indicate that
Table 5, three of the independent varia- other words, if the information search product needs and expectations discon-
bles—needs, expectations, and expecta- tools are perceived to be difficult to use firmation both had a significant effect on
tions disconfirmation—were removed and support services are perceived to be product satisfaction. Although others
from the model. Needs disconfirmation poor, but the user acquired the needed have discussed the importance of needs
alone accounted for 45 percent of the and expected information product, the as a comparison standard when forming
variance in the dependent variable ( F = work for which it was sought can proceed satisfaction judgements,31 currently there
65.72, P < .001; b = .67, P < .001). These with confidence. From a practical per- is little empirical evidence within library
results suggest that needs disconfirmation spective, this finding would seem to sug- sciences research to support this propo-
explains unique variance in the dependent gest that more resources should be sition. This is an important contribution
measure and may be a better predictor of allocated to developing the quality of to models of satisfaction formation pro-
satisfaction with information system/serv- the information products versus the infor- cesses. As noted by Spreng et al.32 and
ices than expectations disconfirmation. mation delivery system. That is, if library others, incorporating needs disconfirma-
users know they will find the information tion into models of satisfaction formation
they need from a library, they will utilize processes eliminates the dilemma embod-
DISCUSSION this library’s resources, even if the infor- ied in the disconfirmation of expectations
This study explored whether disconfirma- mation delivery system or services are model. That is, if a library user expects
tion theory could explain the satisfaction perceived as poor (e.g., difficult to use, an inferior information product and

128 The Journal of Academic Librarianship


receives such, using a purely expecta- the satisfaction measures. This computa- use the system effectively. Thus, our con-
tions-based model of satisfaction predicts tion of satisfaction obscures the possibility struct ‘‘information system/service satis-
the user will be satisfied. However, when that any single attribute (e.g., accuracy of faction refers to the users’ satisfaction
the user’s needs are also taken into the information) may be a more important with attributes of the information system
as well as satisfaction with the information
consideration, the prediction changes. determinant of satisfaction when com- services received.’’
When needs as well as expectations are pared to the other attributes (e.g., precision 2. E. R. Cadotte, R. B. Woodruff, & R. L.
considered, for satisfaction to be high the of the information or relevance of the Jenkins, ‘‘Expectations and Norms in Mod-
information product must exceed both information). els of Consumer Satisfaction,’’ Journal of
the user’s expectations as well as the Third, the present study did not in- Marketing Research 24 (August 1987):
user’s needs. From a practical perspec- vestigate how various situational varia- 305 – 314.
tive, the results of our research indicate bles may moderate the relationships 3. Jagdip Singh, ‘‘A Multifacet Typology of
that fulfilling the library users’ needs contained in the disconfirmation model Patient Satisfaction with a Hospital Stay,’’
around information products plays an of CS/D. Examples of the situational Journal of Health Care Marketing 10 (4)
important role in shaping customers’ variables for information consumers that (December 1990): 8 – 21.
4. D. Halstead, D. Hartman, & S. L. Schmidt,
feelings of satisfaction. Thus, understand- may affect the proposed relationships ‘‘Multisource Effects on the Satisfaction
ing library user needs and delivering include how important the sought infor- Formation Process,’’ Journal of the Acade-
information products that meet these mation is to the user or how urgently my of Marketing Science 22 (2) (1994):
needs is important. the information is needed. Future re- 114 – 129.
Interestingly, disconfirmation of user search should investigate how situation- 5. M. E. Murfin & M. G. Gugelchuk, ‘‘Devel-
needs and not expectations predicted sat- al variables may impact the proposed opment Testing of a Reference Transaction
isfaction with information system/serv- relationships. Assessment Instrument,’’ College & Re-
ices. This finding, that need disconfir- search Libraries 48 (1990): 314 – 348.
mation predicts services satisfaction, is IMPLICATIONS AND CONCLUSIONS 6. R. Applegate, ‘‘Models of User Satisfac-
consistent with several past research find- The findings of this investigation pro- tion: Understanding False Positive,’’ Re-
search Quarterly 32 (4) (1991): 525 – 540.
ings in the satisfaction area.33 Several vide valuable suggestions for the field 7. B. Thompson, C. Cook, & F. Heath, ‘‘A
researchers have suggested that disconfir- of library and information science. First, Short Form of the LibQUAL+(TM) Proto-
mation standards may vary depending on the findings support the disconfirmation col Assessing Users’ Perceptions of Li-
the product or service category. For ex- paradigm as applicable when studying brary Service Quality,’’ Library Quarterly
ample, Oliver34 argued that for continu- satisfaction formation processes in li- (in press).
ously provided products or services such brary users. Second, the findings support 8. A. Bhattacherjee, ‘‘Understanding Informa-
as cable television or public utilities, the notion that library user’s needs and tion Systems Continuance: An Expecta-
customer expectations become passive expectations function as disconfirmation tion – Confirmation Model,’’ MIS Quar-
and are not actively processed. Parasura- standards that influence satisfaction with terly 25 (3): 351 – 370; A. Martensen &
man et al.35 argue that the evaluation the information product and satisfaction L. Gronholdt, ‘‘Improving Library Users’
Perceived Quality, Satisfaction and Loy-
process for services (as opposed to prod- with information delivery system or alty: An Integrated Measurement and Man-
ucts) may be different, that is, based on services. However, needs disconfirma- agement System,’’ Journal of Academic Li-
different sources and types of expecta- tion, rather than expectations disconfir- brarianship 29 (3) (May 2003): 140 – 147.
tions or based on the evaluation of pro- mation, appears the most robust predic- 9. R. L. Oliver, ‘‘A Cognitive Model of the
cess as well as outcome. Future research tor of library user satisfaction. Third, Antecedents and Consequences of Satis-
should explore disconfirmation processes overall library user satisfaction appears faction Decisions,’’ Journal of Marketing
for pure services versus services bundled more a function of the user’s satisfaction Research 17 (November 1980): 460 – 469;
with products. with the information product itself rather G. A. Churchill & C. Surprenant, ‘‘An
than satisfaction with the information Investigation into the Determinants of
Consumer Satisfaction,’’ Journal of Mar-
Limitations of the Current Study system/services accessed to obtain the
keting Research 19 (4) (November 1982):
information product. In conclusion, we 491 – 504.
This study surveyed faculty and admin- believe this study is a solid step in 10. G. A. Churchill & C. Surprenant, ‘‘An
istrators working in academic institutions. providing a theoretical framework to Investigation into the Determinants of
Academic libraries (including remote ac- guide practical information management Consumer Satisfaction,’’ Journal of Mar-
cess to the institutional resources) were activities. keting Research 19 (4) (November 1982):
used as the research context. Therefore, 491 – 504; D. K. Tse & P. C. Wilton,
the information-searching activities stud- ‘‘Models of Consumer Satisfaction For-
ied in this research project were restricted NOTES AND REFERENCES mation: An Extension,’’ Journal of Mar-
to academic settings only. In addition, our 1. Library users may (1) interact directly with keting Research 25 (2) (May 1988):
sample was nonrandomly selected. As a the information system without assistance 204 – 212.
result, generalizability of the results from from library staff; (2) interact indirectly 11. E. W. Anderson & M. W. Sullivan, ‘‘The
this study may be limited to information with the information system through the Antecedents and Consequences of Custom-
assistance of the library staff; or (3) inter- er Satisfaction for Firms,’’ Marketing Sci-
consumers in similar settings.
act directly with the information system ence 12 (2) (1993): 125 – 143; R. L. Oliver,
Second, in this study, information prod- ‘‘A Cognitive Model of the Antecedents
while obtaining assistance from service or
uct satisfaction and information service support staff. The last condition is analo- and Consequences of Satisfaction Deci-
satisfaction were measured as the aggre- gous to users who access an online infor- sions,’’ Journal of Marketing Research 17
gate of their attribute components. Thus, mation system but who request online help (November 1980): 460 – 469; J. C. Olson &
all attributes were given equal weight in from the technical services support staff to P. A. Dover, ‘‘Disconfirmation of Consum-

March 2004 129


er Expectations through Product Trial,’’ tion,’’ Communications of the ACM 26 (10) 32. R. A. Spreng, S. B. Mackenzie, & R. W.
Journal of Applied Psychology 64 (2) (1983): 785 – 793. Olshavsky, ‘‘A Reexamination of the De-
(1979): 179 – 189; D. K. Tse & P. C. Wil- 21. J. Baroudi & W. Orilowski, ‘‘A Short-Form terminants of Customer Satisfaction,’’
son, ‘‘Models of Consumer Satisfaction Measure of User Information Satisfaction,’’ Journal of Marketing 60 (3) (July 1996):
Formation: An Extension,’’ Journal of Journal of Management Information Sys- 15 – 33.
Marketing Research 25 (2) (May 1988): tems 4 (4) (1988): 45 – 59. 33. J. B. Barbeau, ‘‘Predictive and Normative
204 – 212. 22. R. A. Spreng, S. B. Mackenzie, & R. W. Expectations in Consumer Satisfaction: A
12. D. Halstead, D. Hartman, & S. L. Olshavsky, ‘‘A Reexamination of the Deter- Utilization of Adaptation and Comparison
Schmidt, ‘‘Multisource Effects on the Sat- minants of Customer Satisfaction,’’ Journal Levels in a Unified Framework,’’ In Con-
isfaction Formation Process,’’ Journal of of Marketing 60 (3) (July 1996): 15 – 33. ceptual and Empirical Contributions to
the Academy of Marketing Science 22 (2) 23. R. A. Spreng, S. B. Mackenzie, & R. W. Consumer Satisfaction and Complaining
(1994): 114 – 129; J. C. Olson & P. A. Olshavsky, ‘‘A Reexamination of the De- Behavior, H.K. Hunt, R.L. Day (Eds.)
Dover, ‘‘Disconfirmation of Consumer terminants of Customer Satisfaction,’’ (Bloomington, IN: Indiana University
Expectations through Product Trial,’’ Journal of Marketing 60 (3) (July 1996): School of Business, 1985): 27 – 32; R. A.
Journal of Applied Psychology 64 (2) 15 – 33. Spreng & R. W. Olshavsky, ‘‘A Needs
(1979): 179 – 189; Jagdip Singh, ‘‘Under- 24. R. A. Spreng, S. B. Mackenzie, & R. W. Congruency Model of Consumer Satisfac-
standing the Structure of Consumers’ Sat- Olshavsky, ‘‘A Reexamination of the De- tion,’’ Journal of Academy of Marketing
isfaction’’ 19 (3) (1991): 223. terminants of Customer Satisfaction,’’ Science 21 (3) (Summer 1993): 169 – 177;
13. J. Baroudi & W. Orilowski, ‘‘A Short-Form Journal of Marketing 60 (3) (July 1996): J. E. Swan & I. F. Trawick, ‘‘Satisfaction
Measure of User Information Satisfaction,’’ 15 – 33. Related to Predictive vs. Desired Expecta-
Journal of Management Information 25. R. A. Spreng, S. B. Mackenzie, & R. W. tions and Complaining Behavior,’’ In Refin-
Systems 4 (4) (1988): 45 – 59; B. Ives Olshavsky, ‘‘A Reexamination of the Deter- ing Concepts and Measures Consumer
& M. H. Olson, ‘‘User Involvement and minants of Customer Satisfaction,’’ Journal Satisfaction and Complaining Behavior,
MIS Success: A Review of Research,’’ of Marketing 60 (3) (July 1996): 15 – 33. H. Keith Hunt & Ralph L. Day (Eds.)
Management Science 30 (5) (1984): 586 – 26. R. A. Spreng, S. B. Mackenzie, & R. W. (Bloomington, IN: Indiana University,
603; B. Ives, M. H. Olson, & J. J. Baroudi, Olshavsky, ‘‘A Reexamination of the De- 1980): 7 – 12.
‘‘The Measurement of User Information terminants of Customer Satisfaction,’’ 34. R. L. Oliver & J. E. Swan, ‘‘Equity and
Satisfaction,’’ Communications of the Journal of Marketing 60 (3) (July 1996): Disconfirmation Perceptions as Influences
ACM 26 (10) (1983): 785 – 793. 15 – 33. on Merchant and Product Satisfaction,’’
14. In the literature, the terms disconfirmation 27. P. G. Patterson, L. W. Jonson, & R. A. Journal of Consumer Research 16 (Decem-
standards, prepurchase comparison stand- Spreng, ‘‘Modeling the Determinants of ber 1989): 372 – 383.
ards, and preconsumption comparison Customer Satisfaction for Business-to- 35. V. A. Zeithaml, L. L. Berry, & A. Parasura-
standards are used interchangeably. Business Professional Services,’’ Journal man, ‘‘The Nature and Determinants of
15. J. B. Barbeau, ‘‘Predictive and Normative of the Academy of the Marketing Science Customer Expectations of Services,’’ Jour-
Expectations in Consumer Satisfaction: A 25 (1) (1997): 4 – 17. nal of the Academy of Marketing Science 21
Utilization of Adaptation and Comparison 28. J. Cohen & P. Cohen, Applied Multiple Re- (1993): 1 – 12; A. Parasuraman, V. A. Zei-
Levels in a Unified Framework,’’ In Con- gression/Correlation Analysis for the Be- thaml, & L. L. Berry, ‘‘Reassessment of Ex-
ceptual and Empirical Contributions to havioral Sciences, 2nd ed. (New Jersey: pectations as a Comparison Standard in
Consumer Satisfaction and Complaining Erlbaum, 1983), p. 70. Measuring Quality: Implications for Further
Behavior, H. K. Hunt & R. L. Day (Eds.) 29. L. F. Pitt, R. T. Watsom, & C. B. Kavan, Research,’’ Journal of Marketing 58 (Janu-
(Bloomington, IN: Indiana University ‘‘Service Quality: A Measure of Informa- ary 1994): 111 – 124.
School of Business, 1985): 27 – 32. tion Systems Effectiveness,’’ MIS Quar-
16. R. A. Spreng & R. W. Olshavsky, ‘‘A terly 19 (2) (1995): 173 – 187. ***
Needs Congruency Model of Consumer 30. Richard L. Oliver & L. Gerald, ‘‘Effect of
Satisfaction,’’ Journal of Academy of Mar- Satisfaction and its Antecedents on Custom- Xi Shi, an associate professor of li-
keting Science 21 (1993): 169 – 177. er Preference and Intention,’’ Advances in brary services, is head librarian at SUNY
17. R. A. Spreng, S. B. Mackenzie, & R. W. Consumer Research 8 (1) (1981): 88 – 93; Rockland Community College. She holds
Olshavsky, ‘‘A Reexamination of the Deter- R. W. Oshavsky & R. A. Spreng, ‘‘A ‘De- a PhD in MIS from Stevens Institute of
minants of Customer Satisfaction,’’ Journal sire as Standard’ Model of Consumer Sat- Technology and MLS and EdM degrees
of Marketing 60 (3) (July 1996): 15 – 33. isfaction,’’ Journal of Satisfaction, from Rutgers University. Combining over
18. F. Heath, M. Kyrillidou, D. Webster, S. Dissatisfaction, and Complaining Behavior thirteen years experience as a librarian
Choudhury, B. Hobbs, M. Lorie, & F. 2 (1989): 49 – 54; R. A. Spreng, S. B. Ol-
with advanced studies in information
Nicholas, ‘‘Emerging Tools for Evaluating shavsky, & R. W. Olshavsky, ‘‘A Reexami-
Digital Library Services: Conceptual nation of the Determinants of Customer management, her research focuses satis-
Adaptations of LibQUAL+ and CAPM,’’ Satisfaction,’’ Journal of Marketing 60 (3) faction formation processes in informa-
Article No. 170, 2003-06-09, Journal of (July 1996): 15 – 33; R. A. Westbrook & M. tion users.
Digital Information 4 (2) (June 2003); G. D. Reilly, ‘‘Value-Percept Disparity: An
S. Choudhury, M. Lorie, E. Fitzpatrick, B. Alternative to the Disconfirmation of Ex- Patricia Holahan, an associate profes-
Hobbs, G. S. Chirikjian, A. Okamura, & N. pectations Theory of Consumer Satisfac- sor of management at Stevens Institute of
E. Flores, ‘‘Comprehensive Access to Print t i o n , ’’ I n A d v a n c e s i n C o n s u m e r
Technology, teaches graduate courses in
Materials (CAPM),’’ Proceedings of the Research, vol. 10, Richard P. Bagozzi &
First ACM+IEEE Joint Conference on A. M. Tybout (Eds.) (Ann Arbor, MI: Asso- the management of technology and orga-
Digital Libraries (Roanoke, VA: ACM ciation for Consumer Research, 1983): nizational design and theory. Her research
Press, 2001): 174 – 175. 256 – 261. focuses on the implementation of new
19. e.g., Bhattacherjee, 2001; Cook & Thomp- 31. e.g., Spreng et al., ‘‘A Reexamination of technology and the management of
son, 2001; Patterson et al., 1997. the Determinants of Customer Satisfac- cross-functional product development
20. B. Ives, M. H. Olson, & J. J. Baroudi, ‘‘The tion,’’ Journal of Marketing 60 (3) (July teams. Dr. Holahan obtained her PhD in
Measurement of User Information Satisfac- 1996): 15 – 33. organizational behavior from Purdue Uni-

130 The Journal of Academic Librarianship


versity’s Krannert Graduate School of a PhD in mathematics and statistics. adopted by the U.S. Army and many
Management. Dr. Jurkat’s work building mathemati- of the NATO allies for vehicle design
cal and statistical models of off-road and evaluation. Dr. Jurkat has over
Peter Jurkat, professor emeritus at vehicle mobility led to the development five dozen publications and invited
Stevens Institute of Technology, holds of the NATO Reference Mobility Model presentations.

March 2004 131

Вам также может понравиться