Sunteți pe pagina 1din 21

The perceptions of library service

questionnaire (PLSQ): the development of


a reliable instrument to measure student
perceptions of and satisfaction with
quality of service in an academic library
Christine Doyle
Head of Studies in Psychology, Trinity and All Saints College of
the University of Leeds, Leeds

This paper describes the development of a reliable, short questionnaire (PLSQ) to


measure student perceptions of and satisfaction with quality of service in an academic
library. This formed part of a wider study to evaluate student awareness and use of
library services and staff support. This included the simultaneous development of a
'factual' questionnaire and two small scale observational studies, backed up with
'market research-type' interviews to evaluate the effectiveness of the 'help desk' and
ease of use of the computer catalogue. A 'triangulation' technique was thus used to
provide indirect evidence that the questionnaire was validly measuring user
satisfaction. The results from the initial trial of the questionnaire are described and
compared with relevant findings from the 'factual' questionnaire and the help desk
study. Though two samples completed different questionnaires, the frequency results
arc either in complete agreement or mutually supportive. The help desk study results
are also supportive, again suggesting that the questionnaire is a valid measure.
Following statistical analyses, the final sixteen item PLSQ emerged which produces
internally consistent responses and which appears to measure satisfaction with staff
helpfulness and physical conditions in the library and perceptions of ability to find
information and knowledge of CD-ROM and other services.

Introduction
emember, that when total resources are limited, any case which
can be made out for libraries and information services must be
.weighed against corresponding arguments in other fields ..." (P.
Channon, Minister for the Arts addressing the 1981 Library Association
Conference).
On the surface this statement merely reiterates the universal truth
that resources are finite. However, at a deeper level this may reflect a set
of assumptions and values which Pollitt1 terms 'managerialism' and
which now pervades the whole of the Public Sector within the UK.
Central to this thinking is the belief that the private sector is more
efficient and cost effective because it is more accountable. Further, the
key to this success is good management. Most professionals in any field
have always been concerned about quality of service but now they are
also required to demonstrate that it is cost effective.
139
The New Review of Academic Llbrarianship 1995
Academic institutions and their libraries have not escaped this
trend. Within higher education, the Jarrett report, on efficiency in
universities, several selectivity exercises performed by the Funding
Councils and, most recently, a detailed quality audit of university
teaching by the Higher Education Quality Council, have required
institutions to pay close attention to their effectiveness and efficiency. As
Davies and Kirkpatrick2 comment, as administrators scrutinise expensive
support services such as the library 'only by demonstrating their cost
effectiveness and willingness to respond to 'customer* needs will
academic libraries succeed in attracting extra funding.' The Follett
Report3 also recommends that academic libraries employ a range of
performance indicators including measures of user satisfaction.
However, there are other important reasons for a concern with
quality. Campbell & Schlectef argue that academic libraries rank next to
the lecture room as important learning settings and Marchant et al?
stressed that the provision of good service in academic libraries enhances
the quality of education. The Florida State University Report6 states 'a
good library is the heart of a good university' and few educators would
wish to dispute this. It follows therefore that scrutiny of the quality of this
service is a crucial part of the drive for high standards and value for
money.
At the same time there are other serious pressures operating on
academic libraries. The recent huge expansion in student numbers within
higher education, largely without commensurate increases in levels of
funding, have placed great demands upon library resources. Moreover, an
increasing trend towards modular degree schemes, 'student-centred'
methods of course delivery and more autonomous student learning have
meant that 'customer' needs have become more diverse and intensive.
Finally, the 'information explosion' and rapid advances in information
technology have increased librarians' educational/teaching function.7
Librarians need to be able to assess the effect of these factors on the
quality of service.
All these factors have led librarians to try to assess the quality of
service that they offer as much to try to improve it as to gain additional
funding. However, defining quality is no easy matter. Obviously the
perceptions of users must form an important part of this assessment and
has long been acknowledged as such, but quantifying this sort of
information is also complex.
'Objective indicators' of a library's performance such as stock
turnover and utilisation are well developed,1 but many would argue that
the key to the quality of a professional service lies in the subjective
judgements of those who use it. The use of surveys and questionnaires as
research tools to collect such information have only recently been applied

140
The New Review of Academic Librarianshlp 1995
to libraries although the need for such research had been identified in the
1930s.9 Consequently, this approach is less well developed than more
'objective' methods. Measuring and interpreting these opinions,
judgements and indications of user satisfaction is even more complex and
difficult than collecting more 'factual' data such as levels of use of
services and number and type of problems encountered.
The measurement of any human characteristic, including
individual attitudes, is technically complex and still very controversial10".
An attitude is generally defined as a relatively stable affective
(emotional) or evaluative (judgemental) disposition towards a specific
person, situation or other entity. It follows that any instrument
(questionnaire) which aims to measure attitudes must give rise to
internally consistent responses and responses which do not change
significantly over short periods of time. In other words, the instrument
must be reliable. In addition it needs to measure that attitude and that
attitude only and to do so accurately. It also needs to discriminate
between people in the sense that it accurately measures the extent to
which different people hold that attitude e.g. who is more positive, who is
more negative. In the jargon of psychological assessment, these latter
characteristics of an instrument constitute its validity and it takes many
years of research to establish this.
The problems are compounded by the fact that attitudes connected
to user satisfaction, for instance, are likely to be multi-faceted. A user
may be very satisified with some aspects of a library and extremely
dissatisfied with others. Cronin & Taylor12 also argue that customer
satisfaction is best defined as the gap between user expectations and
actual assessments of service delivery. Satisfaction can also vary widely
as a function of specific good or bad experiences of using a service.
Oppenheim10 provides a simple but useful framework for thinking
about what an instrument may be measuring when he discusses attitudes
in terms of a tree analogy. He describes the relationships and patterns of
connections among the different levels of constructs associated with
attitudes in terms of the twigs, branches, trunk and roots of a tree. At the
most superficial level of twigs, we have perceptions and beliefs about
various aspects of the service. At the 'deeper' level of branches and limbs
we have attitudes which influence these perceptions and beliefs. At a yet
deeper level of the trunk we have basic values which may concern issues
such as the meaning of academic study to the individual and the role of
the library in this. Finally, at the deepest level of the roots there may be
something akin to personality. These constructs become more enduring
and pervasive in their influence the deeper the level.
However, we should not underestimate the importance of beliefs.
In the job satisfaction literature, for instance, recent approaches propose

The New Review of Academic Librarianship 1995 141


that subjective, socially constructed realities and associated emotions are
what determine satisfaction, not objective events.1314 In the case of an
academic library where a close knit community constantly use the service
over a period of three, four or more years such socially constructed
realities are likely to be far more important than in most 'customer'
situations which may involve a single transaction and no immediate
'community'. If this is the case, the very concept of 'customer
satisfaction' is inappropriate and too simplistic for academic institutions.
Be that as it may, if Oppenheim's model is anything approaching the
truth of the matter, beliefs and perceptions will reflect much deeper
processes which contribute to these 'socially constructed realities.'
Thus we must be cautious in our interpretation of what an
instrument designed to elicit the opinions of users is actually measuring.
The present study aimed to begin the development of a reliable and valid
'attitudinal-type' questionnaire which would highlight students'
perceptions of and beliefs about library services and in the process to try
to identify the main causes of satisfaction or dissatisfaction. A statistical
technique known as factor analysis may reveal some deeper influences on
these beliefs. Specifically, it aimed to produce a short questionnaire
which could be used in the future to make a quick assessment of student
perceptions. However, while some indirect information was collected to
assess its validity, many more years of research would be needed before
this can be fully established.

The current study


Background
The library staff of Trinity and All Saints College of the University of
Leeds (TAS) commissioned the Psychology department to conduct an
evaluative study of student awareness and use of library services and
staff support to indicate if and where improvements were needed. The
development of the 'attitudinal-type' questionnaire (PLSQ) formed part
of this larger study.
The College had 1638 full time undergraduates enrolled at the
time of the study, of whom 67% were women and 33% men. These
students followed joint degree programmes composed of a choice of one
of several academic disciplines and one of three professional courses.
Only the 1638 undergraduates enrolled at the College were included in
the study.
The Library had at that time a staff of eleven full time equivalents
(FTE) including four FTE professional librarians. It has a collection of
over 130,000 items of multi-media materials and provides computer
facilities linked to the College network, including CD-ROM databases.
Shortage of space for the collection and for individual study was known

142
The New Review of Academic Librarianship 1995
to be a major problem at the time of the survey although the building of
an extension is now underway.

Choice of methodology
All social science research methodologies have their advantages and
disadvantages. The problems and pitfalls of using questionnaires to
evaluate opinions and attitudes are well documented both in the
psychology and information science literatures (see e.g. Oppenheim,10
Saal & Knight," Line,9 Tauber & Stephens,16 Blagden17).
Observation studies are often advocated as an alternative to
questionnaires but they also suffer from various limitations. (See e.g.
Wilson & Streatfield" for a discussion of some of the issues in a library
context.)
More objective indicators of a library's performance, for instance
the effectiveness of inputs such as staff time and new acquisitions in
relation to measurable outputs such as stock turnover and levels of use,
also have their adherents (e.g. Ford,'). However, such measures give little
insight as to the reasons why a particular service appears to be successful
or, more importantly, why some are under-used. Most such measures can
only indicate the effectiveness of administrative systems and procedures.
The TAS staff already collect such data via various computerised systems
and for instance, the 'help desk' staff, record, for example, the number
and type of queries during their periods of duty. Thus two major
objectives of the study were:
i) to collect data that would complement statistics already available as
management information and
ii) to provide a more subjective picture of students' perceptions of
library services and levels of use of particular facilities.
In addition the research aimed to:
iii) identify groups of students who make no use or minimal use of the
library and to discover why;
iv) identify particular areas of difficulty which students encounter in
using the library;
v) to refute or confirm the impressions gained by librarians as to the
difficulties and misconceptions that prevent students from making
effective use of library services.
To these ends the present study employed a 'triangulation'
technique which involved a combination of research methods and
effectively four inter-related studies. These were:
i) A survey questionnaire primarily designed to elicit 'factual'
information regarding such issues as user characteristics (course,

The New Review of Academic Librarianship 1995


143
year, gender etc.). awareness and use of the range of facilities
available, frequency of library visits, difficulties encountered, the
effectiveness of information skills training received (its timing,
helpfulness, etc.), the usefulness of guides and handouts provided by
library staff etc.
ii) The development of a short reliable 'attitudinal-type' instrument
(PLSQ) that would identify the main factors underlying student
satisfactions and dissatisfactions with library services which would
allow a quick check of user perceptions in future years. This was
particularly directed at objectives i), ii), iv) and v) above and
supplements the information obtained from the 'factual*
questionnaire.
iii) Two small scale observational studies, backed up by 'market
research-type' interviews to evaluate the effectiveness of the 'help
desk' and ease of use of the computerised catalogue.
To the extent that the results of the four studies were consistent
the findings of the two main questionnaire instruments would be partially
validated. In other words if the findings from all four studies were in
broad agreement, one could have confidence that user opinions were
being accurately and reliably assessed. However, it must be remembered
that in an ideal world all participants would have completed both
questionnaires on two occasions separated in time. Both instruments
would also have been 'cross-validated' by being completed by a new
sample of participants. To the extent that it was not possible to do this
within the confines of the present study, this research must be regarded as
a pilot investigation.
This paper gives the results for the attitudinal-type instrument and
is aimed at readers who are not familiar with the technical literature of
psychological assessment. Relevant findings from the main 'factual'
survey and the 'help desk' observation study are also referred to as
appropriate.

Method
An item pool of 72 questions was compiled after exhaustive discussions
with librarians and interviews with thirty student users. Each question
was composed of a short statement (e.g. 'It's very important to have
access to a good library when you are studying for a degree', 'I can
concentrate on my work in the TAS library') followed by a five point
scale of response alternatives ranging from strongly agree to strongly
disagree. The questions concerned five aspects of the library which had
emerged as being important during the preliminary discussions. These
were: general attitudes to libraries and academic study; opinions of

144 The New Review of Academic Librarianshlp 1995


physical/working conditions in the library; perceptions of the ease of
locating materials and information; opinions of the helpfulness and
approachability of library staff; opinions of the adequacy of the
collection. To check on the internal consistency of responses and
response biases (e.g. people who agree with everything regardless) the
same questions were often worded in different ways or were presented in
both positive and negative formats. Though most questions were
straightforward and did not try to hide their purpose, a few indirect
questions such as 'I wish students weren't so selfish in their use of library
materials' were also included in this initial trial.
This questionnaire was piloted on a small sample of 35 students
who were then interviewed about their reactions. The data was then
subjected to a preliminary item analysis which helps to identify the most
informative and reliable questions by means of statistical techniques.
Eventually this resulted in a thirty item questionnaire which was then
given its main trial with the entire undergraduate body during one week
in November 1993.
Distributing the questionnaire to some 1600 students, especially
when a second 'factual' survey questionnaire was being developed
simultaneously, posed considerable practical problems. When
questionnaires are left for undergraduates to complete and return at their
discretion response rates tend to be low. 2 " Moreover, leaving the
questionnaires in the library itself would not have reached those who
rarely use it and who may have the most negative opinions of its services.
Thus the questionnaires were distributed at lectures where entire year
groups were taught simultaneously. This meant that the conditions under
which the questionnaires were completed could be reasonably controlled
and the researchers could ensure that respondents gave their own
opinions and did not confer with others. However, the need to minimise
disruption meant that students could not be asked to complete both
questionnaires (the 'factual' survey and the prototype PLSQ) so copies of
each were interleaved and distributed simultaneously. In this way, which
questionnaire was completed by any individual depended entirely on
which seat was being occupied. There is thus no reason to suppose that
the samples completing each questionnaire differed in any systematic
way from each other. Analyses of the major characteristics of each sub-
sample (e.g. gender, year group, course etc.) support this assumption.
However, to begin to establish the validity of the instrument the
responses of each individual to both questionnaires would need to be
compared. Thus if the findings from the two questionnaires in this study
are in agreement this can only be indicative and not definitive evidence
of reliable and valid measurement. Further, practical constraints such as
time and anonymous participation (designed to elicit greater honesty)

145
The New Review of Academic Librarianship 1995
meant that it was impossible to administer the instrument twice and
compare the stability of responses over the two occasions. Is it necessary
to bear these limitations in mind when considering the results.

Following inspection of the frequencies of degrees of agreement


and disagreement with each question, the data was again treated as an
item pool and subjected to an item analysis. The much larger sample
made this procedure much more appropriate and a further fourteen items
were found to be unreliable or unrelated to any general factor and were
dropped from the analyses. This left the data for sixteen questions which
was then analysed using a statistical technique known as factor analysis.
This technique allows us to identify the key dimensions of users
perceptions' of the library services and gives us some idea of what the
instrument is actually measuring.

The sample
812 undergraduates were sampled in the main study which included both
questionnaires. This represents 50% of the total number of first degree,
full-time students enrolled at the College. While this represents a very
good response rate for this population it says little for attendance rates at
lectures! One possible source of bias in the sample could therefore be that
it contains a preponderance of the more conscientious students. However,
the total sample and both sub-samples contained representatives from all
year groups, academic courses and professional courses. The gender
balance also reflected the total enrollment at the College which is two
thirds female. The characteristics of the total sample and the sub-sample
(N=417) of people who completed the 'attitudinal-type' instrument
(prototype PLSQ) are given in table 1,.
(Please note that the percentages in this table are of the total
number of students enrolled in each grouping).
Education students are under-represented compared with other
professional course groups because of difficulties in reaching all the
different groups associated with these courses. As a result, academic
courses which can be combined with education are also under-
represented. There is also a smaller proportion of Year 1 students
compared with the other year groups. However, both the total sample and
the sub-sample appear to be sufficiently representative of the student
body to provide useful findings.
Results Section 1: Prototype PLSQ: Frequency data and
comparisons with the findings from the 'factual' survey questionnaire
(All thirty items).

146 The New Review of Academic Librarianship 1995


TABLE 1: Characteristics of the total sample and sub-sample

Characteristic Frequency
Total Sample Sub-sample

Male 277 51 140 26


Female 525 48 273 25
Yearl 233 36 119 19
Year 2 252 57 132 30
Year 3 223 59 112 30
Year 4 95 55 52 30

Professional Course
Education 206 33 111 18
Public Media 362 63 180 31
Business management 230 54 121 28

Academic Course
Business Studies 36 51 19 27
Cultural Studies 77 67 35 30
Economics 42 51 23 28
English 77 51 36 24
French 54 68 29 38
Geography 45 40 30 27
History 76 55 30 22
Home Economics 26 43 17 28
Maths 32 34 20 21
PE & Recreation 81 59 39 28
Psychology 65 48 36 26
Science & Technology 19 27 11 16
Sociology 56 56 27 27
Spanish 40 57 22 15
Theology 45 31 23 16

Note: to reduce the burden of numerical data on the reader the


strongly agree/agree categories of responses are collapsed as are the
disagree/strongly disagree categories. Neither agree nor disagree
responses are ignored and percentages are rounded up or down in the
tables which follow.

i) General attitudes to libraries and academic study


The frequency results for items believed to tap this aspect of library use
are given in table 2 below.

The New Review of Academic Librarianship 1995 147


TABLE 2: Percentage frequencies of agreement and disagreement for items
intended to measure general attitudes to libraries and academic study

Item % Agree % Disagree

Ql It's very important to have access to a good


library when you're studying for a degree 98 1

Q71 can get by with my degree studies without


having access to a good library 9 85

Q14 You can get a good degree without using the


library much 9 78

Q15 When I'm preparing course work it really


matters to me that I can find the information
I need in TAS library 86 5

Q24 I don't much like reading academic books 26 41

Q29 I usually find something interesting in the


books I borrow from TAS library 52 14

These results are internally consistent. They suggest that the vast
majority of the sample regard the library and its facilities as central to
their degree studies. They are also consistent with the results of the
'factual survey', in which 96% of the sample said that they visited the
library at least a few times a month and 83% visit more than once a week.
There is a small percentage who admit that they do not enjoy reading
academic books and believe they can prepare good coursework and get a
good degree without recourse to library resources. This may support the
interpretation given to the 'factual' survey results which provided
evidence that the main factor influencing frequency of use may be how
conscientious individual students are. Everyone feels that the library is an
important resource but some are more enthusiastic about using it than
others.
However, the very clarity of the results for these items counts
against them since they tend not to discriminate between individuals
(almost everyone gives the same responses) so they do not appear to be
sensitive measures of general attitudes to libraries and academic study. It
seems highly likely that 'social desirability' is influencing these
responses. In other words, people are tending to give the responses they
think they ought to give rather than ones that truly reflect what they
think, feel and do.

148 The New Review of Academic Librarianship 1995


ii) Opinions about the physical/working conditions in the library
The results for items thought to tap this aspect of the quality of library
provision is given in table 3.

TABLE 3: Percentage frequencies of agreement/disagreement with items


intended to measure opinions regarding the physical/working conditions in the
library

Item % Agree % Disagree

Q21 can concentrate on my work in


the TAS library 27 42

Q81 like working in the TAS library 18 41

Q201 find I can't concentrate on my work


in the TAS library 43 21

Q25 I dislike working in the TAS library 31 28

These results are again internally consistent. They suggest that


only a minority of students find conditions in the library conducive to
academic work. The validity of these findings is supported by data from
the 'factual' survey questionnaire. These latter findings include:
• Only half the sample report having used the library for reading and
study.
• The most popular reason given for library visits is to borrow and
return materials (81%) and there are high rates of photocopier use
(64%).
• 56% report that they often or sometimes cannot find a space in the
library to work.
• 61% report that the library is often or sometimes too noisy to
concentrate.
Together, these findings indicate that students tend to borrow
materials to work at home and that, among other things, noise and
overcrowding lead to widespread dislike of working in the library. These
results most probably reflect the great increases in student numbers over
the past few years which have placed great pressure on existing
resources. A new extension to the TAS library, currently being built, and
which will include designated areas for 'noisy' or group work, may help
to alleviate this situation.

Hi) Perceptions of the ease of locating materials and information


Many of the questions in the 'factual' survey questionnaire were devoted

The New Review of Academic Librarianship 1995 149


to exploring students' information retrieval skills and the problems they
encountered when trying to locate information and materials in the
library. The results for the items intended to tap this in the 'attitudinal-
type' instrument (prototype PLSQ) are given in table 4.

TABLE 4: Percentage frequencies of agreement/disagreement with items


intended to measure ease of locating materials and information

Item % Agree % Disagree

Q3 I can usually find the information I


need in TAS library 32 42

Q91 find that I waste a lot of time trying


to find the information I need in the
TAS library 54 18

Q101 wish I knew more about all the


services available in TAS library . 57 13

Q16 Usually I can find the book I want on


the shelves of TAS library 25 53

Q171 feel well informed about how to use the


CD-ROM databases available in TAS library 23 62

Q211 find using the CD-ROM databases very


useful in helping me to find information
on particular topics 36 22

Q261 feel well informed about all the


services available in TAS library 20 49

Q301 find using the CD-ROM databases


frustrating because you rarely get the
information you're looking for 26 24

From this it is evident that a need for further information about


library services, and information retrieval methods, especially about how
to use CD-ROM databases, is required. 42% disagree that they can
usually find the information they need and over half the sample feel they
waste time locating information, wish that they knew more about the
services available and experience difficulties in locating materials on the
shelves. Most appear to be very uncertain about the use of CD-ROM
databases.
All this is entirely consistent with the findings from the 'factual'
survey questionnaire which include the following:
150
The New Review of Academic Librarianship 1995
• Up to 20% of the sample appeared to be unaware of some major
services and awareness and use of some information retrieval
methods such as bibliographies, abstracts and CD-ROM was very
much lower than this
• 84% of the sample report problems in locating materials on the
shelves even after they have been identified in the catalogue
• 72% report that they cannot find the information they need as a
frequent or occasional problem
• 39% report that they cannot find journals and periodicals on the
shelves and 45% found it more difficult to locate relevant information
in journals than in books
55% of respondents said that they would probably or definitely attend
'advanced' library skills training courses if they were available in the
later years of their course and this percentage increased from 52% in
year 1 to 76% in year 4.
• 73% of the sample said that leaflets explaining the use of CD-ROM
databases would be useful.
Thus even though these results involve different people
completing different questionnaires they are mutually supportive. We can
have confidence, therefore, that students genuinely experience these
problems and perceive these as serious shortcomings in the service
provided by the library.

iv) Perceptions of the approachability and helpfulness of library


staff
TABLE 5: Percentage frequencies of agreement/disagreement with items
intended to measure perceptions of the helpfulness of library staff

Item % Agree % Disagree

Q4 Frequently, the TAS library staff are not


very helpful when I have a query 26 45

Ql 1 The TAS library staff are usually friendly


when I ask them a question 61 16

Q18 The TAS library staff are usually very well


informed and can help me 47 18

Q22 The TAS library staff are usually very


helpful when I have a query 51 16

Q27 Sometimes the TAS library staff don't know


enough themselves to be able to help me 28 33

151
The New Review of Academic Llbrarianshlp 1995
In general the majority of students perceive staff to be helpful,
friendly and well informed but a minority seem to feel that this aspect of
the library service could be improved. The 'factual' survey questionnaire
did not attempt to address this issue but a few questions have some
relevance. For instance, asking library staff for help in beginning a
literature search was found to be more effective than asking lecturing
staff. Library induction tours were judged to be successful in helping
people to find their way around and at informing them of the services
available, by the majority of those who took part in them.
More direct corroboration of these findings comes from the results
of the small scale observation study of the readers enquiry desk. Forty
undergraduates were observed as they interacted with the librarian at the
help desk and this was immediately followed up with a 'market research-
type' interview. Of these, 33 (89%) experienced no difficulty in getting
their query answered, 34 (92%) felt the librarian had been helpful, 28
(76%) had their query answered immediately, 34 (85%) felt that the
librarian had been approachable and 29 (73%) rated the help they had
received as good or very good.
From this it appears that those staffing the desk are providing a
high quality service but it is unclear whether all library staff are equally
effective. Perhaps given the small number of fully qualified librarians
and the considerable deficits in information retrieval skills that students
seem to have, these findings taken together are much more positive than
might be expected. They suggest that, in general, staff relate warmly and
professionally towards users and that this is appreciated.

v) Opinions of the adequacy of the collection


The questions intended to tap this aspect are necessarily diverse and most
were found to be unreliable in the first item pool test, probably because
this aspect constitutes a multi-faceted concept. The results for the six
'surviving' items used in the main study are given in table 6.
The results from the 'factual' survey questionnaire indicated that
students had problems in obtaining the information they needed and these
results support these findings. Over half agree that there is too much old
material in the collection with only a tiny percentage disagreeing with
this. In the 'factual' survey 44% said that not enough recent books on
topics of interest was a frequent problem.
55% agree that they often have to go to other libraries and this is a
much higher proportion than the 27% who reported in the 'factual'
survey that they had visited another library. However, also in the
'factual' survey, 80% said that they often or sometimes found that what
they wanted was not in the catalogue and this would indicate a need to
visit other libraries.
152
The New Review of Academic LIbrarianship 1995
TABLE 6: Percentage frequencies of agreement/disagreement with items
intended to measure opinions of the adequacy of the collection

Item % Agree % Disagree

Q5 There are too many old outdated books in


the TAS library 57 9

Q61 often have to go to other libraries


because I can't find the information I
need in the TAS library 55 20

Q12 The TAS library should spend more of


its budget on things like journals 28 19

Q131 wish other students weren't so selfish


in their use of library materials 57 12

Q23 What I want in TAS library is books,


not all these other materials 26 46

Q28 Lecturers shouldn't be allowed to have


books on loan for so long 60 9

57% wish that other students were not so selfish in their use of
materials and 60% begrudge lecturers their long loan periods. This agrees
with the 'factual* survey findings which include the following.
• 80% said that what they wanted was often or sometimes out on loan
• 72% said that they often or sometimes could not find the information
they needed.
Only 28% agree that more money should be spent on journals and
the majority are undecided on this issue. Again the 'factual' survey
results support the validity of these findings in that only a minority of
students read journals at all frequently or feel that their ability to prepare
coursework would be seriously affected if they were cancelled. Large
numbers also felt that journals were difficult to locate and use. Given all
this, it is hardly surprising that students are unclear about whether buying
more journals is a good idea.
However, the corresponding question on spending more money on
books attracts more disagreement than agreement and this appears to be
inconsistent with what has just been said about journals. Perhaps students
feel that more money should be spent on everything and have no strong
preferences regarding priorities.

Conclusion
Although two samples completed different questionnaires, the frequency

153
The New Review of Academic Llbrarlanshlp 1995
results are either in complete agreement or mutually supportive. We can
thus have some confidence that the findings are valid.

Results Section 2: The construction of the PLSQ: a short, reliable


questionnaire
This phase of the data analysis involved three main steps:
i) treating the thirty item questionnaire as an item pool and conducting
an item analysis;
ii) performing a statistical procedure known as factor analysis to identify
the main dimensions of opinions underlying responses;
iii) calculating reliability coefficients to assess the internal consistency of
each person's responses.
The first step identified fourteen items which were either
unreliable or unrelated to any underlying dimension and were dropped
from the analysis. This left sixteen items which had good internal
consistency and the data from these was then entered into the factor
analysis.
Factor analysis is a controversial, notoriously unstable and highly
complex procedure although it is simple enough to perform using suitable
software. As Child20 notes, 'a central aim of factor analysis is the 'orderly
simplification' ... of a number of interrelated measures ... to make order
out of the apparent chaos of the environment.* He gives a helpful analogy
from the field of medicine. By noting clusters of symptoms which always
seem to appear together doctors can infer some kind of underlying
disease. However, the picture is complicated by the fact that symptoms
overlap. A high temperature, for instance is a symptom of many different
illnesses.
Factor analysis attempts to do something similar for human
characteristics but these have an additional problem in that they cannot
be observed and measured with anything like the accuracy and precision
of something like body temperature. We thus have to exercise great
caution when interpreting our findings.
Child devotes a whole chapter in his book to defining a factor but
a provisional idea can be gained from considering the relationship
between someone's overall height and the length of his or her arms and
legs. If we correlated these three measures we would probably find a
strong relationship — tall people will tend to have longer limbs than
short people. Thus we might say that underlying overall height, leg
length and arm length is one factor which we might call linear size. Thus
we have simplified three measures to one factor which underlies all three.
Other measures which we might have included in the analysis, such as
eye colour or left-handedness would be unlikely to show any relationship
with linear size and so would not appear in this factor.

154 The New Review of Academic Librarianship 1995


In the same way, if responses to a cluster of items are all
intercorrelated with each other and have little relationship with other
items then it may be that there is an underlying factor. The strength of the
relationship of any one item to the underlying factor is assessed by a
statistic known as a factor loading which is similar to a correlation. These
can range from 0.0 (no relationship) to plus or minus 1.0 (a perfect
positive or negative correlation). Thus the closer the factor loadings are
to 1.0 the stronger that item's relationship to the underlying factor.
However, since our measurement of characteristics like opinions and
attitudes is necessarily inaccurate, loadings of above 0.3 are considered
minimally adequate and above 0.7 as quite good evidence that the factor
actually exists (though giving it a name remains something of a 'black
art').
Further, we can have more confidence that the factors actually
exist if the same ones emerge when we use different methods of factor
analysis on the data. Four different methods of doing factor analysis were
performed on the data for the sixteen 'surviving' items and the same four
factors emerged from all the analyses. These appear to be:
i) Satisfaction with the helpfulness of staff
ii) Satisfaction with physical conditions in the library
iii) Perceptions of ability to find information
iv) Perceptions of knowledge of CD-ROM and other services.
Together these four factors explained 57.6% of the variance in
responses. The items and their factor loadings (principal components
with oblimin rotation method) are given in table 7.
Please note that these items were not presented in blocks like this
on the questionnaire but were randomly intermixed. (See tables 2-6 for
question numbers.)
There is some evidence that there are relationships between the
factors themselves. Factors 1 and 2 are negatively correlated, though the
relationship is rather weak (the correlation is -.26). This indicates that the
more favourably students perceive staff helpfulness to be the less positive
they are about CD-ROM and the less well informed they feel about
library services. This is somewhat puzzling but may be because students
who are unsure about services are more likely to have approached
librarians with queries and have thus experienced generally helpful
responses. Perceptions of staff helpfulness is linked to students' liking for
working in the library (correlation =.24) and their perceptions of their
ability to find information (correlation =.39). Thus it appears that the
perceived helpfulness of staff is an important influence on general
satisfaction with the library.

155
The New Review of Academic Llbrarlanshlp 1995
TABLE 7: Factor loadings for the 16 item questionnaire

Item Factor loading

Factor 1
The TAS library staff are usually very well
informed and can help me .85
The TAS library staff are usually very helpful
when I have a query .83

Sometimes the TAS library staff don't know


enough themselves to be able to help me .71

Frequently, the TAS library staff are not very helpful


when I have a query .70

Factor 2
I like working in the TAS library .79

I dislike working in the TAS library .78

I find I can't concentrate on my work in

the TAS library .72

I can concentrate on my work in the TAS library .72

Factor 3
I can usually find the information I need in
the TAS library .76
Usually I can get the book I want from the shelves
in TAS library .74
I often have to go to other libraries because I
can't find the information I need in TAS library .67

I find that I waste a lot of time trying to find the


information I need in the TAS library .57

There are too many old, outdated books in the


TAS library .54

Factor 4
I feel well informed about how to use the CD-ROM
databases available in the TAS library .84

I find the CD-ROM databases very useful in


helping me to find information on particular topics .72

I feel well informed about all the services


available in TAS library .69
156
The New Review of Academic Librarianshlp 1995
However, the strongest relationship found is between satisfaction
with the physical conditions in the library and perceived ability to find
needed information (correlation =.56). Although we cannot say that one
causes the others, this does indicate that those who are satisfied with one
aspect of the library are also satisfied with other aspects whilst those who
are dissatisfied are likely to be dissatisfied with a range of things.
These results are sufficiently clear for us to have reasonable
confidence about what the sixteen item questionnaire is measuring but an
unreliable instrument can never be valid. It would be the equivalent of
trying to measure someone's height with a piece of elastic!
A statistic which is often used to assess how internally consistent
people's responses to a questionnaire are, is called Cronbach's Alpha.
This measures the extent to which all the items are measuring the same
thing — in this case we could argue it is general satisfaction with library
services. Cronbach's Alpha is based on correlation and like correlations,
the closer to 1.0 the more reliable the instrument. Cronbach's Alpha for
the sixteen item questionnaire as a whole was 0.81 — a very respectable
level — which means that it is reliable enough to be able to reach
sensible conclusions about groups of people. This means that if further
investigation shows the new questionnaire to be valid (and the signs are
hopeful) then it can be used to make a quick assessment of general levels
of student satisfaction with library services.
Finally, when each person's scores on all the items are added
together and presented in the form of a frequency histogram (Figure 1)
we find that the overall 'satisfaction' scores for this sample are
reasonably normally distributed. This means that the majority of peoples'
scores are clustered on either side of the mean (average) with fewer and
fewer people scoring towards the extremes. For various technical reasons
this is also a good sign. In particular, it shows that the questionnaire is
discriminating between people in terms of their levels of general
satisfaction. The average score for this sample is 50.85, 13% of the
sample are seriously dissatisfied and 15% are pretty well satisfied with
their academic library.

Summary and conclusions


The results for the 'attitudinal-type' questionnaire (prototype PLSQ) are
supported by the findings from both the 'factual' survey questionnaire
and the small scale observation study of the library help desk. This gives
indirect evidence of its validity.
The sixteen item PLSQ is internally reliable. It appears to be
measuring satisfaction with staff helpfulness and physical conditions in
the library and perceptions of ability to find information and knowledge
of CD-ROM and other services.

157
The New Review of Academic Librarianship 1995
25.0 30.0 350 40.0 45.0 60.0 SS.0 80.0 66.0 70.0

Score
FIG. 1: Frequencies of total scores

Scores for the sample are normally distributed and the instrument
appears to be discriminating between different people's levels of
satisfaction.
The PLSQ may seem rather limited in its scope but it does seem to
tap six of the nine dimensions underlying quality of library service
identified by Davies and Kirkpatrick2 in their study. It may be that this
latter instrument could be a useful tool to help to establish its validity.
The sixteen item PLSQ now needs to be cross-validated on a new
sample and administered twice to the same sample to establish the
stability of responses over time. Further investigations of its validity also
need to be undertaken. However, there are grounds for optimism that this
will prove to be a useful quick measure of general levels of student
satisfaction, at least within the institution for which it was designed.
It would be most interesting to discover whether the PLSQ would
'transfer' to other academic libraries and prove equally reliable and
produce the same underlying dimensions. Much more work on its
validity and stability is also needed. Researchers who are interested in
using it in their work are requested to contact the author.

References
1. POLLITT, C. Managerialism and the public services, 2nd edition. Blackwell, 1993.
2. DAVIES, A. and KIRKPATRICK, I. To measure service: ask the library user. Library
Association Record, 96 (2), 1994, 88-89.
3. Joint Funding Councils' Libraries Review Croup: Report (Follett Report). Bristol:
HEFCE, 1993.
4. CAMPBELL, D.E. and SCHLECTER, T.M. The University Library: an important
setting for the study of environment-behaviour relationships. Man-Environment
Systems, 8 (1), 1978, 41-42.
158
The New Review of Academic Librarianship 1995
5. MARCHANT, M.P., BROADWAY, M.D., ROBINSON, E. and SHIELDS, D.M.
Research into learning resulting from quality school library media service. School
Library Journal, 30, 1984, 22-24.
6. Florida University Steering Committee of the University Self Study. The Future of
Florida State University. Tallahassee: Florida State University, 1962.
7. MANORAMA, S. The organisational climate of university libraries. Library
Management, 14 (1), 1993,28-30.
8. See, for example, FORD, G. Approaches to performance measurement: some
observations on principles and practice. British Journal of Academic Ubrarianship, 4
(2), 1989, 74-87.
9. LINE, M.B. Library surveys: an introduction to the use, planning, procedure and
presentation of surveys. London: Clive Bingley, 1982.
10. See, for example, OPPENHEIM, A.N. Questionnaire design and attitude measurement.
London: Heineman, 1966, and
11. BEECH, J.R. and HARDING, L. (eds) Testing people: a practical guide to
psychometrics. NFER/NELSON, 1990.
12. CRONIN, J.J. and TAYLOR, S.A. Measuring service quality: a re-examination and
extension. Journal of Marketing, 56, July 1992, 55-68.
13. LANDY, F.J. The psychology of work behaviour (3rd edition). Dorsey Press, 1985.
14. BANDURA, A. The social foundations of thought and action: a social cognitive
theory. Prentice Hall, 1986.
15. SAAL, F.E. and KNIGHT, P.A. Industrial/occupational psychology: theory and
applications. Brooks/Cole, 1989.
16. TAUBER, M.F. and STEPHENS, I.R. Library surveys. New York: Columbia
University Press, 1968.
17. BLAGDEN, J. Do we really need libraries? Proceedings of the First Joint Library
Association, Cranfield Institute of Technology Conference on Performance and
Assessment. Cranfield Press, 1983.
18. WILSON, T.D. and STREATFIELD, D.R. Structured observation in the investigation
of information needs. Social Science Information Studies, 1, 1981, 173-184.
19. MALLEY, I. Aspects of user education in UK academic libraries: 1976-1981.
Education Libraries Bulletin, 24 (3), 1981, 1-15.
20. CHILD, D. The essentials of factor analysis. 2nd edition. Cassel Educational Ltd.,
1990.

Acknowledgements
This project was conducted as part of the final year teaching programme
of the Class of 1990 psychology students who contributed to the initial
questionnaire design and did much of the questionnaire administration,
scoring and data entry. In return they were led through the complexities
of the process of developing a reliable and valid instrument to measure
human characteristics. The whole formed an extremely worthwhile
project.
I must also thank the staff of the TAS library, especially Liz
Murphy, Gillian Leach and the Director, Marlene Godfrey, who provided
courteous and thoroughly professional help and assistance throughout.
They also gave helpful comments on an earlier draft of this paper though
its deficiencies remain entirely my own.
159
The New Review of Academic Ubrarianship 1995

S-ar putea să vă placă și