Professional Documents
Culture Documents
Ojsadmin,+Journal+Manager,+ART Lorenzetti
Ojsadmin,+Journal+Manager,+ART Lorenzetti
Based Library and Information Practice 2007, 2:1
Evidence Based Library and Information Practice
Feature Article
Identifying Appropriate Quantitative Study Designs for Library Research
Diane Lorenzetti
Centre for Health and Policy Studies
University of Calgary
Calgary, Alberta, Canada
E‐mail: dllorenz@ucalgary.ca
Received: 02 December 2006 Accepted: 12 February 2007
© 2007 Lorenzetti. This is an Open Access article distributed under the terms of the Creative Commons
Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use,
distribution, and reproduction in any medium, provided the original work is properly cited.
Abstract
This paper is concerned with the identification of quantitative study designs suitable for
library research. Identifying a researchable question and selecting a research method best
suited to it are key to the successful design and execution of any research project. Each
research situation is unique, and each researcher must find the method that best suits both
their situation and the question at hand. Following a brief discussion of issues related to
question development, the author outlines a checklist that may assist the process of
selecting study designs for quantitative research projects. When faced with options in terms
of study design selection, pragmatic issues such as expertise, funding, time, and access to
participants may influence this decision‐making process.
Introduction some light on research design selection by
Identifying a researchable question and introducing a means by which researchers
selecting a research method best suited to it can match their questions to a selection of
are two important decisions that any appropriate study designs. The focus of this
investigator must make when embarking on paper will be on quantitative research.
a research project. It is these processes with
which this paper is concerned. What is quantitative research?
Research designs may be qualitative,
After a preliminary discussion of question quantitative, or mixed methods techniques
development, this paper attempts to shed
3
Evidence Based Library and Information Practice 2007, 2:1
which employ both qualitative and An investigation into whether or not
quantitative methods. researchers who receive training in
literature searching skills are more
Qualitative research focuses on exploring successful in obtaining funding from
the views of individuals. “Observing events granting agencies than those who do not is
from the perspectives of those involved” an example of a research problem that could
(Powell and Connaway 3) is used as a be answered through an experimental
means to develop theories to explain events research design. Here, the intervention is the
or behavior (Creswell, Educational Research; training and the outcome or effect is the
Powell and Connaway). Because a number of successful grant applications.
qualitative research approach is exploratory,
it is often used when a topic is new or little True experimental designs are those in
understood, or where “existing theories do which research participants can be
not apply” (Creswell, Research Design 22). randomly allocated to receive or not receive
an intervention of interest. Quasi‐
Quantitative research, on the other hand, is experimental designs are experiments in
primarily concerned with testing theories which researchers, for whatever reason, are
and measuring the relationships between not able to randomly assign participants and,
variables or the impact these variables have instead, conduct research with naturally
on outcomes of interest (Couchman and occurring groups ‐ for example, students
Dawson; Creswell, Educational Research). registered in an introductory political
Whereas qualitative researchers use events science course (Creswell, Educational
and individual experiences as a way of Research).
developing theories, quantitative
researchers begin with a hypothesis or Developing a researchable question
theory and then try to prove or disprove it. Before tackling the demanding task of
Generally speaking, a quantitative approach launching a research project, we should ask
might be recommended if a research project ourselves the following question: What do
involves the identification of “factors that we want to know?
influence an outcome, the utility of an
intervention or understanding the best In a survey published in a 2001 issue of
predictors of outcomes” (Creswell, Research Hypothesis, medical librarians were asked to
Design 21‐22). Within this broad category “identify the most relevant research
are a variety of research designs, including questions currently facing the profession”
surveys, bibliometrics, and experimental (Evidence Based Librarianship
designs or techniques. Implementation Committee 9). The variety
and complexity of these responses clearly
Experimental techniques is a category of indicates that the profession continues to
quantitative research design wherein the grapple with many challenges. Among other
researcher attempts to determine the impact things, we are concerned with issues of
of an intervention (e.g., a teaching technique, resource selection, budgeting, library skills
electronic database, or collection training, the adoption of new technologies,
development policy) on an outcome or effect and the need to determine what impact our
of interest. An example of an experimental profession has on the clients we serve
design is a randomised controlled trial. In (Evidence Based Librarianship
experimental research, investigators are Implementation Committee). There appears
interested in exploring issues of “cause and to be no lack of potentially researchable
effect” (Creswell, Educational Research 283). issues of interest to our profession.
4
Evidence Based Library and Information Practice 2007, 2:1
An awareness of a given topic and its clearly stated (Booth; Richardson and
potential options or issues provides Wilson). An example of a foreground
researchers with the information they need question might be: “Is Blackboard an
to decide on the scope and direction of their effective tool for online library literacy
research. Those unfamiliar with aspects of a course delivery?”
given topic will find it useful to explore
their area of interest prior to developing PICO is a model that can help facilitate the
their research questions. A literature review development of a researchable question.
may identify whether or not research
already exists in librarianship or a related PICO stands for:
field. Conducting quality original research P population
can be time consuming. It may be that I intervention(s) or exposure
adequate evidence already exists that can be C comparison
adapted to your needs. O outcome(s) of interest
The importance of developing specific Population is the group under
questions that accurately reflect research observation/study (in the case of
needs cannot be overstated. collection use studies, the population
could well be the library’s collection of
Background versus foreground questions books, journals, or other media)
Questions may be loosely grouped into two
types: background and foreground. When Intervention/Exposure is the action or
you need to gather together basic intervention of interest
underlying background knowledge on a
topic, you are dealing with a background Comparison is an alternative action or
question (Booth; Richardson and Wilson). In intervention against which the
addressing a background question, you intervention of interest will be measured.
attempt to arrive at a greater understanding
of the complexity and diversity of the Outcome is a measure of the desired
literature as a whole, either as an endpoint effect of the Intervention/Exposure
in itself or as a first step towards developing
a more focused foreground question. An Although PICO was developed as a clinical
example of a background question might be: model, it can be adapted to other disciplines
“What online systems exist to facilitate the of research, including librarianship.
delivery of distance education courses?”
So why use PICO? Identifying the PICO or
Now imagine that you have identified building blocks of a question/problem can
through research that Blackboard (an online identify gaps in question development and
tool that allows students and instructors to may inspire investigators to rethink and/or
communicate over the Internet) may be an reevaluate an original premise, making for a
effective tool for the delivery of distance solid researchable question.
education courses in your institution. You
would like to determine if it would be Using the PICO model may also point the
effective by comparing Blackboard to your way towards specific research designs
existing method for online course delivery. amenable to particular types of questions of
A foreground question is very focused in its which yours is but one. For instance, if it
research aim. It contains multiple appears that you will be comparing one
components, which are clear and can be intervention to another, this identifies your
5
Evidence Based Library and Information Practice 2007, 2:1
question as belonging to a family of addition to students or perhaps some subset
questions which can be effectively of the student population).
researched by employing a certain type of
quantitative study design. On the other Identifying the research method
hand, if it appears that the research question Once a researchable question has been
properly belongs to that category previously identified, the next step is to match it to an
identified as a background question, this too appropriate research design. Each design
is valuable information that may inform the has its own strengths and weaknesses.
choice of research design. While some may be ideal for tracking trends,
others are better at demonstrating cause and
So how can this be used? Imagine we are effect or predicting outcomes. The nature of
evaluating the reference services we offer to a research question should guide the design
students. We would like to know if an in‐ selection. At the same time, there may be
person, drop‐in reference service would more than one research design that could be
absorb more or less librarian time than an appropriate for your project. As with
email reference service. question development research design
selection can be, in some instances, an
If we apply the PICO model, the iterative process.
components of this question might be
identified in this way: In a 2002 article on levels of evidence in
evidence based librarianship, Eldredge
Population: students identified three categories of research
Intervention/Exposure: in‐person drop‐in questions: prediction, intervention and
reference service exploration (ʺEBL levels of evidenceʺ).
Comparison: email reference service Prediction questions anticipate outcomes;
Outcome: frequency of use intervention questions compare actions or
processes; and exploration questions may be
The research question that results from this equated with background questions in that
exercise could be: they often describe or synthesize
observations or results (Booth; Eldredge
Would an in‐person drop‐in reference service be ʺEBL levels of evidenceʺ). Eldredge
used more frequently by students than an email organized specific research designs
library reference service? according to their appropriateness for
answering one of these three question types
Although this example fits neatly into the (ʺEBL levels of evidenceʺ). This paper was
PICO format, in many instances no used as the basis for developing the
comparison exists in which case “No checklist that follows for the selection of
Intervention/Exposure” (or placebo) is the quantitative research designs.
comparison.
Using the checklist, librarians new to the
Rarely are questions born in their final field of research may identify possible
format – they almost always require some research designs based on the focus or goal
tweaking. At this point you might decide to of the question under consideration.
re‐evaluate your question to focus on any Consider, in general terms, what you hope
other outcomes of interest (e.g., amount of to achieve through your research. Determine
time spent responding to each request) or which of the activities and corresponding
populations (e.g., university faculty in study designs is conducive to the aims or
goals of your research.
6
Evidence Based Library and Information Practice 2007, 2:1
You may find that more than one research Descriptions of Research Designs
design fits your needs. When deciding on an
approach, consider not only the suitability Bibliometrics
of the research design but the expertise, Bibliometrics is a method for describing
resources, time and other variables that will patterns of publication within a body of
be integral to the successful completion of literature. It has been used in librarianship
this piece of research. In other words, what to identify core literatures, classify
resources will you need in order to research literatures, predict publishing trends,
your question in the manner you have describe patterns of book use, and chart the
determined to be the most effective? If those dissemination of ideas (Powell and
resources are not available, a different Connaway 63).
design may need to be selected.
Checklist for Selecting a Quantitative Study Design from Eldredge, “EBL Levels of Evidence”
Match the desired aims of your research activity to one or more of the activities and
corresponding research designs outlined below. Descriptions of the research designs follow the
checklist.
Report on observations and analyze experiences with a process, service, program, technology or
individual:
Case Study (aka Case Report)
Identify trends in attitudes, behaviors or characteristics where:
Participants will be surveyed at one moment in time
Cross Sectional Survey
Participants will be surveyed at multiple points in time
Longitudinal Survey
Identify patterns of publication within a given field or body of literature:
Bibliometrics/Citation Analysis
Predict whether the presence or absence of an exposure or event influences an outcome of
interest where:
Participants will be observed from just prior to the exposure into the future
Prospective Cohort Study
Participants will be identified from historical records and followed into the present
Retrospective Cohort Study
Demonstrate if the presence or absence of an intervention, service, or process causes an outcome
of interest. Here the investigator will attempt to test the strength of cause and effect relationships
where:
7
Evidence Based Library and Information Practice 2007, 2:1
No comparison/control group is available and the intervention group will act as its own
comparison via before and after tests
Single Group Controlled Comparison Study
Participants cannot be randomly assigned to intervention or comparison/control groups and pre‐
existing groups must be used
Controlled Comparison Study
Participants can be randomly assigned to intervention or comparison/control groups
Randomized Controlled Trial
Participants may be followed over time and multiple observations (pre‐intervention, during the
intervention, and post‐intervention) may be made
Time Series
Map key concepts and/or identify gaps in a research area:
Scoping review
Evaluate existing published evidence with respect to a specific research question:
Systematic Review
Pool or combine the results of published studies to determine a measure of effectiveness:
Meta‐analysis
Calculate the probability of expected consequences of implementing different strategies:
Decision Analysis
Citation analysis is a form of bibliometrics in techniques. Eldredge reported in 2004 that
that it focuses on identifying patterns of case studies ranked with surveys and
citation (Powell and Connaway 63). Citation qualitative methods as one of the three most
analysis can be used to identify seminal used research designs in librarianship (ʺEBL:
studies in fields of research, trace the an overviewʺ 294). Case studies are a means
citation history of studies, or identify of reporting on in depth observations and
influential writers within a field of research. explorations of an individual, population,
process, technology, or service (The
For more information on bibliometrics, Cochrane Collaboration Secretariat;
readers are advised to consult Powell and Creswell, Research Design 15; Eldredge,
Connaway’s Basic Research Methods for ʺInventoryʺ; Fraenkel and Wallen). No
Librarians. baseline information is collected and no
comparison groups are used. For these
Case Study reasons, case studies are not capable of
Although case studies/case reports rank low proving cause and effect … although they
on most evidence hierarchies, they employ can begin the chain of evidence that leads to
few resources, are relatively easy to a realization of why or how something
undertake, and for these reasons, are often happens (Eldredge, ʺInventoryʺ).
the first layer of evidence published with
respect to emerging questions, issues, or Citation Analysis ‐ See Bibliometrics
8
Evidence Based Library and Information Practice 2007, 2:1
Cohort Study outcomes will reflect, for example,
A cohort is a population of individuals who differences in the makeup of the two groups
share a characteristic (Ruane 99). In a cohort rather than differences in the interventions)
study a naturally occurring group (the by attempting to closely match
cohort) is followed over time it is exposed or characteristics of participants in the
not exposed to a particular event. The intervention and comparison/control groups
outcomes of exposed or non exposed (Eldredge, ʺEBL: an overviewʺ 295). In the
members of the cohort are compared case of the instructional techniques example,
(Eldredge, ʺCohort studiesʺ; The Cochrane we might want to ensure that members of
Collaboration Secretariat). A cohort study is both groups are in the same year and
an observational study and researchers do program of study, are comparable in age
not manipulate the cohort in any way. They and background, and have had similar past
simply observe the course of events in an exposure to library instruction. Conducting
attempt to predict whether or not there may both pre‐tests and post‐tests also helps
be a relationship between the event and an account for variations in these naturally
outcome or outcomes of interest (Streiner occurring groups. Controlled comparison
and Norman). For this reason, cohort studies are referred to as a type of quasi‐
studies are sometimes referred to as event‐ experimental design.
based designs (Ruane). Prospective cohort
studies follow participants “into the future” For more information on controlled
(The Cochrane Collaboration Secretariat 8); comparison studies, readers are advised to
whereas retrospective or historical cohort consult Creswell’s Educational Research
studies identify participants “from historical Methods and Powell and Connaway’s Basic
records and follow them from the time of Research Methods for Librarians both of
those records into the present” (The which contain detailed chapters on
Cochrane Collaboration Secretariat 8). experimental and quasi‐experimental
designs.
For more information on cohort designs,
particularly as they relate to librarianship, Cross sectional Survey ‐ See Survey
readers are advised to consult Eldredge’s
2002 article on this topic (ʺCohort studiesʺ). Decision Analysis
Decision analysis is a transparent decision
Controlled Comparison Study making method wherein all options and
In controlled comparison studies, two possible consequences of each choice are
groups are compared (Eldredge, ʺEBL:an made explicit. It “involves identifying all
overviewʺ 295). Comparisons are made available alternatives and estimating the
between an intervention group receiving probabilities of [and assigning values to]
intervention A and a comparison /control potential outcomes associated with each
group B receiving either intervention B or alternative (Evidence Based Medicine
no intervention (Fraenkel and Wallen 273). Working Group 409). Existing research and
The aim is to determine if these data are used to develop the model and
interventions (e.g., two different options, and the potential results of those
instructional techniques) have an effect on options are represented graphically as a
an outcome of interest (e.g., knowledge decision tree. Researchers use the
retention, course grades, etc…). Participants probability and value of all outcomes to
are not randomised to either group. Instead, develop a “quantitative estimate of the
controlled comparison studies attempt to relative merit of [each alternative]”
reduce bias (the chance that observed (Evidence‐Based Medicine Working Group).
9
Evidence Based Library and Information Practice 2007, 2:1
For more information, readers are advised opportunities for bias that exist in other
to consult Keeney’s overview on this topic. research designs. That being said, it is not
always possible or appropriate to randomise
Longitudinal survey ‐ See Survey participants. For instance, imagine that your
library has a subscription to a particular
Meta‐analysis electronic database. It is expensive and you
Meta‐analysis is the “use of statistical are wondering if it is worth retaining. It may
techniques in a systematic review to be difficult (not to mention unethical) to
integrate the results of included studies” randomise students to either have or not
(The Cochrane Collaboration Secretariat 25). have access to a database or other service for
By statistically pooling the results of which there is already system‐wide access.
multiple studies, it is possible to determine, On the other hand, the benefits of email
with greater certainty, the effect or non‐ versus in‐person reference service, one
effect that an intervention has on an process of handling document delivery
outcome of interest (Egger, Smith, and requests versus another, one‐time versus
Altman; Powell and Connaway). multiple follow‐up library instruction
training, etc … may be successfully explored
Although all meta‐analyses should begin through an RCT design (Eldredge, ʺThe RCT
with a systematic review, not all systematic designʺ).
reviews do or should include a meta‐
analysis. Although it is possible to conduct a For more information on RCT design and
meta‐analysis with as few as two studies, a application, readers are advised to consult
misleading conclusion can result if Eldrege’s 2003 article on this subject
researchers attempt to combine the results of (Eldredge, ʺThe RCT designʺ), Powell and
different study designs (e.g., RCTs and Connaway’s Basic Research Methods for
cohort studies) or include data that is not Librarians, and Creswell’s Educational
comparable. Research: Planning, Conducting, and
Evaluating Quantitative and Qualitative
For more information on meta‐analysis Research.
techniques, readers are advised to consult
Systematic Reviews in Health Care: Meta‐ Retrospective Cohort Study ‐ See Cohort Study
Analysis in Context.
Scoping Review
Prospective Cohort Study ‐ See Cohort study A scoping review is similar to a systematic
review in that it employs rigorous methods
Randomised Controlled Trial to rapidly yet systematically locate and
A randomised controlled trial (RCT) is “an review literature appropriate to the research
experiment in which two or more question under review. Unlike systematic
interventions…are compared by being reviews though, scoping reviews tend to
randomly allocated to participants” (The address broader, less focused questions,
Cochrane Collaboration Secretariat 37). The include a variety of study designs, and tend
intervention group receives the intervention to not assess the quality of included studies
of interest to the researchers and the (Arksey and OʹMalley). Scoping studies are
comparison/control group receives either no often undertaken to map key concepts in a
intervention (ie placebo) or an alternate research area and/or identify gaps in the
intervention (Creswell, Educational existing literature. They can be used to
Research 295). The ability to randomise quickly chart complex areas of research,
participants eliminates many of the provide a visual picture of the range of a
10
Evidence Based Library and Information Practice 2007, 2:1
field, and are also useful in identifying Quantitative, and Mixed Methods
topics where further research might be Approaches (155).
warranted.
For more information on survey designs,
For more information on scoping reviews, readers are advised to consult Fraenkel and
readers are advised to consult Arksey and Wallen’s How to Evaluate Research Design
O’Malley’s article on this topic as it provides in Education.
both an overview and detailed method for
proceeding with this type of review. Systematic Review
A systematic review is “a review of the
Single Group Controlled Comparison Study evidence on a clearly formulated question
A single group controlled comparison study that uses systematic, explicit [and
(also known as a single group pretest reproducible] methods to identify, select
posttest design) is a controlled comparison and critically appraise relevant primary
study wherein only one group is under [original] research, and to extract and
observation by researchers. The group is analyze data from the studies that are
measured or observed both before and after included in the review” (Centre for Reviews
exposure to the intervention of interest. and Dissemination 4). Unlike traditional
Unlike other controlled comparison studies, narrative reviews, systematic reviews are
the intervention group acts as its own comprehensive, unbiased, and systematic in
comparison or control (Creswell, Research the ways in which literature is located,
Design; Creswell, Educational Research; selected, abstracted, and reported.
Fraenkel and Wallen; Powell and
Connaway). For an excellent overview of the process of
conducting systematic reviews, readers are
Survey advised to consult the 2001 report produced
Surveys are “questionnaires administered to by the University of York’s Centre for
a sample of a population in order to identify Reviews and Dissemination entitled
trends in attitudes, opinions, behaviors or Undertaking Systematic Reviews of
characteristics” (Creswell, Educational Research on Effectiveness.
Research 52). Cross sectional surveys or
studies are referred to as snapshots because Time Series
they essentially capture information from a Controlled comparison studies and
group of respondents at one moment in time randomised controlled trial designs assume
and are useful for describing current that the researcher will observe/measure
situations and conditions (Ruane 93). participants at a single point in time (twice if
Longitudinal surveys, on the other hand, a pre‐test is included in the research design).
gather information at multiple points in If you are able to observe participants over a
time and may be used to track changes in period of time, you may wish to consider a
values, opinions, attitudes, and behaviors randomised or non‐randomised time series
(Ruane 95). Longitudinal surveys may approach. In a time series, multiple pre‐test
repeatedly collect information from the and post‐test measures or observations are
same individuals or generate new survey taken both before and after the intervention
samples each time the survey is (Creswell, Educational Research; Polit,
administered (Creswell, Research Design; Tatano‐Beck, and Hungler 177). The
Fraenkel and Wallen; Ruane). A checklist for strength of the time series design is in its
designing a survey method can be found in ability to track the impact of an intervention
Creswell’s Research Design: Qualitative, over time.
11
Evidence Based Library and Information Practice 2007, 2:1
Discussion questions may be more effectively
In a statement released in 2001, the Special addressed through qualitative research
Libraries Association (SLA) commented that techniques, and these approaches should be
“both the need to make decisions quickly considered when choosing an appropriate
and the lack of a clear connection between design. For a discussion of qualitative
much library and information science research designs in librarianship, readers are
research and the day‐to‐day problems faced advised to consult Eldredge’s 2004
by librarians make seeking and applying inventory of research designs (Eldredge,
our own knowledge base a challenge.” ʺInventoryʺ) and Powell and Connaway’s
Although it may be argued that librarians Basic Research Methods for Librarians.
continue to experience a certain lack of Many researchers are also embracing the
professional support with respect to need to incorporate both qualitative and
engaging in research activities, we cannot quantitative methods in their research
claim to be the only profession that is beset projects (Creswell, Research Design).
with tight timelines when it comes to
making and implementing decisions. Conclusion
This paper has touched on the subject of
For those of us not schooled in research research design in only the most cursory
methods however, undertaking a research sense. No attempt has been made to
project of any sort may appear to be a very comprehensively represent the range of
intimidating experience. Certainly a brief research designs that exist and may be
overview article of this sort will, at best, applied or adapted to answering all
only whet the appetite for more information. questions facing our profession. Readers are
Although there is much that can and should encouraged to “research” the process more
be read on this subject before embarking on thoroughly, and speak with colleagues
a research project, it is often in the before proceeding to engage in a research
application that we learn the most. endeavor.
Partnering with an experienced professional Although the focus of this paper has been
colleague may be the best way for a on select quantitative techniques, it is not
budding researcher to become familiar with the intention of the author to recommend
the research process. If this is not an option, the application of one study design or group
we may wish to consider looking a bit of designs over another in all research
further afield. Many granting agencies look situations. Not every question can or should
favorably on proposals submitted by be answered by a randomised controlled
multidisciplinary teams. Librarians have trial. Each research situation is unique, and
skills that make them valuable partners in a each researcher must find the method that
variety of research endeavors. Those of us best suits both their situation and the
who work in academic libraries are question at hand. When faced with options
surrounded by research faculty experienced in terms of study design selection, it may be
in leading multifaceted research that pragmatic issues such as professional
investigations, some of whom may be open expertise, time, and access to participants
to mentoring a beginning researcher in will influence the decision‐making process.
exchange for active participation on their
teams. Sometimes the ideal study design from a
research question perspective is not ideal
Insofar as research designs are concerned, from a practical standpoint; an important
the author wishes to reiterate that many characteristic of a researcher should be the
12
Evidence Based Library and Information Practice 2007, 2:1
ability to recognize and creatively adapt to 2007<http://www.york.ac.uk/inst/crd/r
the challenges of reality. eport4.htm>.
Works Cited Couchman, Wendy and Jane Dawson.
Nursing and Health‐Care Research: a
The Cochrane Collaboration Secretariat. Practical Guide. 2nd ed. London:
ʺGlossary of Terms in the Cochrane Scutari Press, 1995.
Collaboration. Version 4.2.5.ʺ The
Cochrane Collaboration. May 2005. 18 Creswell, John W. Research Design:
Feb. 2007 Qualitative, Quantitative, and Mixed
<http://www.cochrane.org/resources/h Methods Approaches. 2nd ed.
andbook/glossary.pdf>. Thousand Oaks, CA: Sage Publications,
2003.
Egger, Matthias, George Davey Smith, and
Douglas Altman, eds. Systematic ‐‐‐. Educational Research: Planning,
Reviews in Health Care: Meta‐Analysis Conducting, and Evaluating
in Context. Ed. M. Egger, G. Davey‐ Quantitative and Qualitative Research.
Smith, and D. G. Altman. London: BMJ 2nd ed. Upper Saddle River, NJ:
Books, 2003. Merrill, 2005.
Evidence‐Based Medicine Working Group. Eldredge, Jonathan D. ʺEvidence‐Based
ʺHow to use a clinical decision Librarianship: an Overview.ʺ Bulletin
analysis.ʺ Centre for Health Evidence. of the Medical Library Association 88.4
2001. 18 Feb. 2007 (Oct. 2000): 289‐302.
<http://www.cche.net/usersguides/deci
sion.asp>. ‐‐‐. ʺCohort Studies in Health Sciences
Librarianship.ʺ Journal of the Medical
Arksey, Hilary and Lisa OʹMalley. ʺScoping Library Association 90.4 (Oct. 2002):
Studies: Towards a Methodological 380‐92.
Framework.ʺ International Journal of
Social Research Methodology 8.1 (Feb. ‐‐‐. ʺThe Randomised Controlled Trial
2005): 19‐32. Design: Unrecognized Opportunities
for Health Sciences Librarianship.ʺ
Booth, Andrew. ʺFormulating Answerable Health Information and Libraries
Questions.ʺ Evidence‐Based Practice Journal 20.S1 (2003): 34‐44.
for Information Professionals: a
Handbook. Eds. Andrew Booth and ‐‐‐. ʺInventory of Research Methods for
Anne Brice. London: Facet Publishing, Librarianship and Informatics.ʺ Journal
2004. 61‐70. of the Medical Library Association 92.1
(Jan. 2004): 83‐90.
Centre for Reviews and Dissemination.
ʺUndertaking Systematic Reviews of ‐‐‐. ʺEvidence‐Based Librarianship Levels of
Research on Effectiveness: CRDʹs Evidence.ʺ Hypothesis 16.3 (Fall 2002):
Guidance for Those Carrying Out or 10‐13.
Commissioning Reviews.ʺ University
Evidence Based Librarianship
of York Centre for Reviews and
Implementation Committee. ʺThe Most
Dissemination. Mar. 2001. 18 Feb.
Relevant and Answerable Research
Questions Facing the Practice of Health
13
Evidence Based Library and Information Practice 2007, 2:1
Sciences Librarianship.ʺ Hypothesis Utilization. 5th ed. Philadelphia:
15.1 (2001): 9‐17. Lippincott Williams & Wilkins, 2001.
Guyatt, Gordon and Drummond Rennie. Powell, Ronald R. and Lynn Silipigni
Usersʹ Guides to the Medical Literature: Connaway. Basic Research Methods for
Essentials of Evidence‐Based Clinical Librarians. 4th ed. Westport, CT:
Practice. Chicago: American Medical Libraries Unlimited, 2004.
Association Press, 2001.
Richardson, W. S. and M. C. Wilson. ʺOn
Fraenkel, Jack R. and Norman E. Wallen. Questions, Background and
How to Design and Evaluate Research Foreground.ʺ Evidence Based Health
in Education. 5th ed. Boston: McGraw Care Newsletter 17 (Nov. 1997): 8‐9.
Hill, 2003.
Ruane, Janet M. Essentials of Research
Keeney, Ralph L. ʺDecision Analysis: an Methods: A Guide to Social Science
Overview.ʺ Operations Research 30.5 Research. Malden, MA: Blackwell
(Sep./Oct. 1982): 803‐38. Publishing Limited, 2004.
Streiner, David L. and Geoffrey R. Norman. Special Libraries Association. ʺSLA research
PDQ Epidemiology. 2nd ed. Hamilton, statement: putting our knowledge to
ON: B.C. Decker, 1996. work.ʺ Special Libraries Association.
Jun. 2001. 18 Feb. 2007
Polit, Denise F., Cheryl Tatano Beck, and B. <http://www.sla.org/content/resources/
P. Hungler. Essentials of Nursing research/rsrchstatement.cfm>.
Research: Methods, Appraisal, and
14