Sie sind auf Seite 1von 6

Doing a Literature Review

Jeffrey W. Knopf, Naval Postgraduate School

S tudents entering a graduate program


often encounter a new type of assign-
ment that differs from the papers they
In general, a literature review has two
key elements. First, it should concisely
summarize the findings or claims that
or not in dealing with a certain problem.
In this context, one might focus, for ex-
ample, on the “lessons learned” from
had to write in high school or as college have emerged from prior research efforts previous efforts to deal with a certain
undergraduates: the literature review on a subject. Second, a literature review problem ~and those lessons learned might
~also known as a critical review essay!. should reach a conclusion about how have been proposed by outside scholars
Put briefly, a literature review summa- accurate and complete that knowledge is; or by practitioners themselves!.
rizes and evaluates a body of writings it should present your considered judg- To return to the first context, review-
about a specific topic. The need to con- ments about what’s right, what’s wrong, ing existing knowledge can itself be the
duct such reviews is by no means limited what’s inconclusive, and what’s missing end goal if one simply wants to ascertain
to graduate students; scholarly research- in the existing literature. In contrast to the current “state of the art” on a partic-
ers generally carry out literature reviews some other ways of surveying a body of ular subject or problem. In this context
throughout their research careers. In a literature, such as an annotated bibliogra- ~as well as the other two!, it is important
world where the Internet has broadened phy, the literature review is a work of not to simply summarize the available
the range of potentially relevant sources, synthesis. For this reason, it is important research, but also to evaluate it critically.
however, doing a literature review can not to simply write a summary list of Such critical analysis should not be ex-
pose challenges even to an experienced what each individual work says, but in- clusively negative; it is also important to
researcher. stead to focus on the body of work identify positive results to take away
In recent years, I have taught a viewed as a whole. from the existing work.
course designed to help students in a Conducting a literature review can Second, a review of existing knowl-
policy-oriented Master’s program draft have several benefits: edge can be a preliminary step in a
thesis proposals. In looking for readings larger research project. Such a literature
to assign to guide their literature re- • It can give you a general overview review is often required for a thesis or
views for these proposals, I discovered of a body of research with which dissertation proposal; it is also frequently
a paucity of appropriate published you are not familiar. an element in proposals for research
sources.1 The vast majority of methods • It can reveal what has already been grants. The most basic reason to under-
textbooks written for students in politi- done well, so that you do not waste take a literature review in this context is
cal science or public policy contain no time “reinventing the wheel.” to make sure the proposed research ques-
discussion whatsoever of the literature • It can give you new ideas you can tion has not already been answered. If an
review.2 Some general methods texts use in your own research. existing study convincingly answers the
contain sections on the literature review • It can help you determine where question you want to address, it is better
~for example, Cresswell 2003; Patten there are problems or flaws in exist- to find out before you get started than
2005!, but these turned out not to be ing research. when you are in the middle of a research
very helpful in meeting the needs of • It can enable you to place your re- project.
the student population I was teaching. search in a larger context, so that Assuming no prior study has solved
Finally, there are a few books devoted you can show what new conclusions your problem of interest, then the pur-
solely to preparing a literature review might result from your research. pose of your proposal’s literature review
~Fink 2005; Galvan 2005; Pan 2004!, is to situate your proposed project in re-
but these were too long to be a viable Three Contexts for Literature lation to existing knowledge. This en-
reading assignment for the course. In Reviews ables you to address the concept of a
the end, I drafted my own “how to” “contribution to knowledge,” which is
handout on doing a literature review. In In general, literature reviews are pro- important because potential advisors and
the hope that my observations might be duced in one of three contexts: A litera- other people who might review a pro-
helpful to others, I have adapted my ture review can be an end in and of posal generally ask of any research pro-
handout for publication here. itself; it can be a preliminary stage in a posal “what is the expected contribution
larger research project; and it can be a to knowledge?” or “what will be the
component of a finished research report. value added of completing this re-
In any of these contexts, a literature re- search?” The goal here is to show that
Jeffrey W. Knopf is visiting associate view can address either theoretical or people who read the final research prod-
professor of National Security Affairs at the practical questions. In academic settings, uct are likely to learn some new or dif-
Naval Postgraduate School (NPS), Monte- review essays most often focus on the ferent information or argument compared
rey, CA. He is the author of Domestic Soci- theories scholars have proposed to ex- to what they would find in existing stud-
ety and International Cooperation: The plain some phenomenon; sample topics ies. In short, a literature review in a re-
Impact of Protest on US Arms Control Policy
might include the causes of terrorism or search proposal provides an overview of
(Cambridge University Press, 1998). This
essay developed out of materials the author the pre-conditions for democratization. A existing scholarship and explains how
prepared for a course in research methods literature review can also be used, how- your proposed research will add to or
that he helped create for a new profes- ever, to determine and assess the practi- alter the existing body of knowledge.
sional Master’s program in Homeland Secu- cal know-how available in regard to Conducting a literature review at a
rity at NPS. which measures are likely to be effective preliminary stage of a research project

PSOnline www.apsanet.org 127


can also be helpful in stimulating your tribution to knowledge; this might be vehicle for disseminating their work.
own thinking. A broad review of existing new factual information, a new theoreti- Scholars are increasingly posting confer-
literature might reveal new theoretical cal proposition, or a new policy pro- ence papers, working papers, and mono-
hypotheses, research methods, or policy posal. In addition, information or graphs on the Internet.5 These postings
recommendations that you want to incor- reasoned argument that changes our de- are often part of a work in progress that
porate in your own research. gree of confidence in an existing belief is has not been published in a book or jour-
Third, a literature review can be a also a contribution to knowledge. This nal article; they represent these scholars’
component of a finished research report. might be new evidence or analysis that most current thinking. For this reason, it
This literature review will generally in- corroborates and thereby strengthens a can be important to search for such work
volve building on and0or revising the particular belief. It can also be evidence to keep a review of existing knowledge
literature review completed at the pro- or analysis that challenges and thereby as up-to-date as possible.
posal stage. Its purpose is to help show casts doubt on a particular point of view. At the same time, the Internet must be
how your final conclusions relate to the If new information or analysis is power- used with great caution. Most academic
prior wisdom about your subject. ful enough, it might convince people that publications go through peer review,
their prior belief was wrong and lead which in most cases helps ensure that the
them to embrace a different perspective. published work meets certain standards
Ways to Frame the When using a literature review to indi- of scholarship. In contrast, anyone with
Contribution to Knowledge cate where proposed research might access to the necessary equipment can
make a contribution to knowledge, there- post anything they want on the Internet.
The literature review is an attempt to fore, it is helpful to think in terms of Many postings are based on little or no
summarize the existing state of knowl- identifying the existing beliefs people research, make no attempt to be unbi-
edge about a subject and, in research have and the level of confidence with ased, and contain factual claims that are
proposals, to frame the proposed which they hold them. This facilitates the questionable. If you use the Internet to
research’s expected contribution to task of showing where additional re- broaden the range of sources consulted
knowledge. Knowledge, in this context, search could make a difference. in a literature review, be sure to consider
does not necessarily mean “Truth” with a carefully whether the items that you find
capital T. Rather, knowledge refers to are credible and meet at least minimal
beliefs, in particular beliefs that some Consider Casting Your Net standards of scholarly research. Look to
individuals have a degree of confidence Widely see whether the authors have provided
in due to study or experience. In the so- their credentials and consider whether
cial sciences and policy research, many The traditional literature review fo- these make them credible sources on the
hypotheses cannot be proven conclu- cuses on books published by academic subject. Also examine whether an item
sively. When reviewing literature, there- presses and articles published in aca- contains documentation of its sources
fore, it is common to refer to the demic journals. For many purposes, and whether these appear to be credible.
“claims” or “arguments” advanced by a these will continue to be the appropriate If your interest is in existing policy pro-
study or school of thought. Hence, a typ- focus. However, on many questions, es- posals or practices, then academic credi-
ical review of existing knowledge identi- pecially those involving a policy dimen- bility may matter less than other
fies the claims made in a literature and sion, actors besides university-based considerations, such as whether the
assesses the strength of the support of- academics might issue relevant reports. source of information is in a position of
fered for those claims. In addition, the development of the In- authority or has inside knowledge; even
It is helpful to think of knowledge as ternet has made it easier to disseminate in these cases, however, you need to
having two elements: what we believe research reports in formats other than in screen Internet postings to weed out
and how strongly we believe it. Further academic publications. This growth in those that lack a valid basis for their as-
research can affect either or both of alternative research producers and out- sertions. Despite the risks, valuable
those elements, either positively or nega- lets for disseminating research makes it sources of research exist beyond tradi-
tively, and any of these results would be advisable to consider a wider range of tional academic books and journals, and
a contribution to knowledge. This is sim- sources when conducting a review. In- it is worth using the Internet to seek
ilar to the logic of Bayesian analysis in deed, because relevant information and these out. At the same time, be sure
statistics. In Bayesian statistics, if one analysis is increasingly found in sources not to limit your search for sources to
believes a statement has a certain proba- other than traditional academic publica- just the Internet as any college or
bility of being true and then obtains ad- tions, it may be more accurate to think university library will have many items
ditional pertinent data, one can revise the of your task as a “review of existing on its shelves that are not available
estimated probability the statement is knowledge” 4 than as a review of litera- electronically.
true using a mathematical formula pro- ture per se.
vided by Bayes theorem. Even where Other entities that might produce re-
such precise quantification is not feasi- search relevant to your topic include Pointers on How to Create
ble, one can attempt an analogous quali- government agencies, international gov- an Effective Review
tative assessment.3 ernmental organizations, non-govern-
This provides a framework for think- mental organizations, think tanks, and First, especially if this is your first
ing about the possible consequences of independent, freelance researchers. Some literature review, read some existing re-
new research. Further research could of their reports are still produced in print view essays to see how other researchers
create a new belief in an area where peo- form and are available through any good have carried out this task.6 Imitate what
ple have no prior knowledge, it could library collection. Increasingly, though, you think works well, and avoid those
alter an existing belief, or it could many of their reports are released elec- things that strike you as ineffective or
change how much certainty people feel tronically and can be found through care- unnecessary. Chances are that some re-
about a current belief. Most obviously, ful searching on the Internet. Traditional view essays will have been assigned in
something brand new is a potential con- academics are also using the Internet as a some of your graduate classes; if so,

128 PS January 2006


begin by re-examining them. You can were interested in learning about research The Nuts and Bolts: What
also ask your instructors or advisors to on deterrence, for example, you could Questions Should the
suggest literature reviews that they be- search for encyclopedias or handbooks of Literature Review Try to
lieve provide good models to follow. social science, of international relations, Answer?
Second, for each research study you or of conflict and violence. If you cannot
read for your review, be sure you can find a relevant source for your area of A literature review summarizes and
succinctly summarize the study’s main interest, consult a reference librarian or evaluates the state of knowledge or prac-
claim. You should be able to describe in be creative in trying different combina- tice on a particular subject. To do this,
a sentence or two the central argument of tions of keywords when searching an most literature reviews must address four
each item you read. It will not always be online library catalogue. tasks or sets of questions. The first two
necessary to include this information, but There are a couple of other likely steps are to determine what each individ-
having an awareness of each study’s sources for summaries of existing re- ual study has examined and what each
overarching thesis will help you compare search that identify contrasting schools of has concluded from its examination. The
different items as you write your review. thought. Academic journals often publish third step involves summarizing the col-
Third, your written review should be review essays that reflect upon one or lective results. To do this, sort the results
selective. When you write the literature more recently published books on a par- into three categories: what the existing
review, it is often not necessary to dis- ticular topic. Identify the journals that studies and reports have in common,
cuss every item you read. The write-up publish regularly on your topic of interest what the studies disagree about, and
should discuss only the studies that have and peruse the tables of contents for the what they overlook or ignore. Finally, the
a direct bearing on the central focus of past few years to determine whether there fourth step is to reach a judgment about
your review or your proposed research. are recent review essays that could help the quality of the literature overall: what
In addition, rather than summarizing the orient you to a body of research. In addi- are the key findings that appear to be
studies in their entirety, the review should tion, theses and dissertations usually con- valid, and where is more work needed?
focus only on the aspects of those studies tain a literature review section or chapter. To elaborate, the first task is simply to
that are relevant for your purposes. Many dissertations become the basis for be clear about what each item you are
Fourth, when you write a literature books, so identifying books published by reviewing was trying to do. For example,
review, do not simply summarize, item freshly minted Ph.D.s is often a fruitful was the work concerned with theory? If
by item, each publication you have read. way to find recent surveys of a field. so, was the goal explanatory, or did it
A literature review should not have the Many graduate schools also deposit cop- have some other objective? If the goal
following structure: paragraph 1 notes ies of dissertations and theses completed was explanation, what was the dependent
that book A says X; paragraph 2 notes by their students with UMI ~formerly variable for the study? How was it con-
that article B says Y; paragraph 3 notes University Microfilms!. It is now possi- ceptualized and operationalized? Ascer-
that book C says Z; etc. ble to search the UMI collection online, taining this information before you
In general, a literature review should and you can buy copies of theses and compare studies will help you determine
impose some intellectual order on dissertations that appear relevant.8 if they were even examining the same
the material. Therefore—as a fifth Sixth, while seeing how others have problem. Sometimes different studies
pointer—it often helps to think about characterized a field of research is help- reach different conclusions because they
grouping individual studies into larger ful, it is essential not to rely on others’ asked different questions or defined the
“camps” or “schools of thought.” One summaries of existing studies. Review phenomenon of interest in different
can do this in terms of different theories articles in specialized encyclopedias or ways.
they propose or defend, different meth- academic journals are a good place to get The second step involves identifying
odological approaches they take, or started, but they cannot substitute for the main argument in each work. Does it
different policies they favor. Often, alter- your own reading. Read for yourself the have a thesis? If so, how strongly does
native views reflect differences in the sources that are most critical for your the study say its findings support the
disciplines or backgrounds of the own interests and draw your own thesis, and what qualifications or reserva-
authors—academics vs. government offi- conclusions. tions does the author report?
cials, psychologists vs. economists, etc. A seventh and final pointer: Get into The third task listed above—sum-
This can stand as another basis for cat- the habit of associating individual au- marizing existing studies in terms of
egorizing schools of thought. If you thors and major camps or points of view three categories—can be especially valu-
group similar studies together, rather than with each other. In academic writing, able in a research proposal. In short, any
discuss three like-minded authors sepa- scholars often use the last name of the body of research can be usefully summa-
rately in three successive paragraphs, you author of a study as a shorthand to refer rized in terms of the following:
can mention all three together in a single to the theory or argument advanced by
sentence such as ‘A, B, and C argue that that author. For example, in International ~1! Areas of consensus or near-
policy X has been ineffective and pro- Relations Theory, Kenneth Waltz was consensus. On some issues nearly all of
pose policy Y instead.’ one of the leading developers of a theory the relevant experts may agree. Such
For any subject where there is already known as “neo-realism.” In writing about conclusions can be either positive or
a substantial body of research, chances this approach, other authors will switch negative; i.e., they can involve beliefs
are that some scholars have already back and forth between referring to about what is true or what works or
sought to classify the research into con- Waltz, to the Waltzian approach, and to what is false or does not work. Areas of
trasting schools of thought. In such cases, neo-realism. Since this has become stan- consensus represent the “conventional
it is a good idea to start by familiarizing dard practice in scholarly writing and wisdom” about a subject.
yourself with existing summaries of the conversation, it is a good idea to get
research. Many fields or sub-fields have used to thinking about each alternative ~2! Areas of disagreement or debate. In
encyclopedias or other reference works camp both in terms of the generic label many cases, there exists information and
that contain short, introductory essays on by which it is known as well as in terms analysis about a topic but no consensus
the research on particular topics.7 If you of the authors identified with that camp. about what is correct. These areas of

PSOnline www.apsanet.org 129


debate usually give rise to the alternative that case, it is easy to show that pro- the use of different methodologies?
“camps” or “schools of thought” men- posed research on the topic would make Do the studies make clear the meth-
tioned above. a contribution to knowledge. More often, odology by which they have reached
however, the gap will be narrower than their conclusions, or are key claims
~3! Gaps. There may be aspects of a this. People will have studied some, but made purely by assertion? Was the
topic that have not been examined yet. not all, aspects of a problem, or they will methodology used an appropriate
These gaps in knowledge might involve have examined a problem using some choice for the question being re-
questions no one has tried to answer, theories or methodologies, but neglected searched, and was it applied
perspectives no one has considered, or others. In this situation, if your goal is to correctly?
bodies of information that no one has fill the gap you identify, your research
attempted to collect or to analyze. proposal would state something like “re- By identifying and comparing the as-
searchers have studied a, b, and c, which sumptions, theories, data, and methods of
Once you have identified where there are related to the problem of X, but they the studies you review, you can pinpoint
is conventional wisdom, where there are have not studied d, which is also relevant the underlying disagreements responsible
debates, and where there are gaps, you to understanding @or solving# X.” for debates in the literature. You can
can use the literature review to describe The fourth task in a review—evalu- then, if you wish, target your own re-
what will be the contribution to knowl- ating the overall state of knowledge on a search on one of the underlying disagree-
edge of the research you are proposing topic—requires a thorough examination ments, which could help resolve an
and why it will be of interest to your of how the answers given by the litera- existing debate. By evaluating each of
intended audience. Your contribution can ture have been produced. In examining these elements critically, you can also
address any or all of these. For example, how people have reached their conclu- show where there are problems or flaws
you might believe there are reasons to sions, consider evaluating the following: in existing studies and then, if you wish,
doubt the conventional wisdom. In gen- target your own research on fixing one or
eral, you should not accept areas of • Their assumptions. If there are dis- more of these problems in the literature.
agreement uncritically. The fourth task agreements, can they be traced to Finally, as noted previously, you can also
noted above—assessing the quality of the different assumptions made by the look for important issues that the existing
literature—includes probing for areas conflicting studies? Are the key as- research has overlooked and frame your
where the existing wisdom is less than sumptions made by the most impor- research as an effort to fill this gap.
conclusive. The literature review can tant studies a plausible basis for
then be used to highlight potential flaws research, or are they so problematic The Problem of Too Few
in the reasoning or evidence related to an that they call into question the rest Sources and the Problem of
area of consensus. This could be used to of the analysis? Too Many Sources
set up proposed research that might chal- • Their logic. If there are disagree-
lenge the conventional wisdom. ments, can they be traced to differ- Students sometimes choose a research
Weighing in on an existing debate is ent theoretical perspectives? Do the topic, such as how to address a new pol-
another possibility. Here, one uses the studies explain the reasoning that icy problem or what can be learned
literature review to show the likely value supports their key conclusions, or about a recent event, because they think
of research that could help judge the rel- are important arguments made no one has yet studied the issue. In such
ative merits of conflicting points of view purely by assertion? Is the reason- cases, students expect that there will not
or that could help point the way to a use- ing that is provided logically per- be any literature relevant to the question
ful synthesis. suasive, or does it contain internal they want to research. It does not pay to
Finally, proposing to fill a gap in ex- contradictions or make a giant leap be too skeptical on this score; you might
isting knowledge is an obvious way to at a key point in the analysis? be surprised at what you find once you
frame the usefulness of a suggested piece What are the most plausible coun- start to search for resources. Even if you
of research. A gap may involve theory, if terarguments or alternative explana- come up empty, however, this is not a
no scholar on a topic has yet considered tions to the main thesis in each wasted effort. If you can report that a
an important theoretical question or a study, and does each study address serious search uncovered no examples of
particular alternative theory ~e.g., al- these adequately? studies that examined your research
though there is a growing body of re- • Their evidence. If there are dis- question, then you have largely demon-
search on the causes of terrorism, agreements, can they be traced to strated that your research will fulfill the
perhaps no one has yet studied what in- the use of different bodies of evi- “contribution to knowledge” criterion for
fluences terrorist decisions about whether dence or to disagreements about the evaluating research proposals ~I say
to target agriculture versus people!. Or a facts? Do the studies provide evi- “largely” because you still have to show
gap may be empirical, if there is a histor- dence to back up their main claims, that the proposed research could produce
ical case or a source of data no one has or are important claims made purely meaningful results!.
analyzed ~e.g., we know a lot about how by assertion? Is the evidence This still leaves the problem of what to
local emergency responders acted on valid—i.e., is it factually accurate discuss in a literature review. The prob-
September 11, but are there useful les- and on point? Has all the relevant lem of too few sources can usually be
sons that could be learned from how evidence been considered, or have solved by thinking in terms of two tiers
emergency personnel dealt with an inci- some obviously relevant cases or ~or circles! of literature. In the first tier
dent in some other locality, for example, bodies of data been overlooked? Is ~or inner circle!, you are concerned with
an anthrax threat phoned in to a local the evidence that has been consid- studies that directly address your own
Planned Parenthood clinic?!. ered representative, or are the cases proposed research question. In the second
The relevant gaps in knowledge can or data selected for study likely to tier ~or outer circle!, you broaden your
be broad or narrow. In some cases, a have biased the results? 9 review to consider publications that are
topic might essentially be virgin territory: • Their methodology. If there are dis- relevant to or overlap some part of your
no one has studied any aspect of it. In agreements, can they be traced to own question, even though they do not

130 PS January 2006


directly address the same point. If there they might not reflect the current state of to your own research interests. It should
is a reasonable body of work in the first knowledge and debate. Instead, consider also evaluate the state of knowledge in
tier, in many cases this will be all that using one or more of the following rules terms of what’s right, what’s wrong,
you discuss in the literature review. You of thumb: what’s an area of uncertainty or debate
would only go outside this inner circle if that cannot be resolved using the existing
there was some specific other publication ~1! Focus on the leading authorities. You research, and what’s missing because no
that proposed a theory, policy proposal, may discover that certain authors or one has yet considered it carefully. To
or research method that you want to studies are cited quite frequently in the create such a review of existing knowl-
apply in your own research project. literature. These are probably considered edge, it helps to ask and answer the fol-
If there is nothing or very little that is key works, so it is a good idea to re- lowing questions:
directly on the same topic as yours, then spond to what they have to say, even if
your literature review will need to con- it means ignoring some less-influential • What questions have the existing
sider some items in the second tier. You studies. publications addressed? What issues
might consider items that have a theoreti- have been neglected?
cal perspective you want to explore in ~2! Focus on recent studies from high- • What are the main conclusions of
your research. Even if no one has applied prestige or high-visibility sources. You existing research? What do the stud-
the theory to your specific question, it is generally want to emphasize the most ies actually argue?
still appropriate to discuss key works that recent research in the field you are re- • What are the points of convergence
have developed the theory so you can viewing. Among recent studies, look in the literature, and what are the
explain why it might provide a good per- especially for those that have been pub- main disagreements? Where dis-
spective for analyzing your topic. lished in a high-prestige outlet: examples agreements exist, what are the bases
It can also be helpful to think in terms include books from the university press of the disagreement?
of analogies. In particular, are there situ- of a highly ranked university or articles • What theories and0or policies
ations or problems that are similar to the in the leading journal in the field in
and0or evidence has the literature
one you want to study, so that research looked at? What potentially relevant
question. Sources that garner a lot of
on those other problems might contain information and alternative theories
attention are also important to evaluate:
relevant ideas? For example, if you were or policies have not been examined?
interested in identifying ways to protect in some cases, for example, it might be • How solid are the conclusions that
crops from agro-terrorism and you could relevant to assess a book on the best- have been reached? Are they based
find no studies directly on this topic, you seller list. on sound reasoning, careful assess-
could consider looking for research on ment of the evidence, and a well-
~3! Focus on the studies that are most
efforts to protect crops against natural executed methodology? Or are there
relevant and helpful for your question of
disease outbreaks. If your literature re- good reasons to doubt some of the
view reveals findings about ways to ad- interest. The more a study is directly on existing conclusions?
dress the latter problem, you could then point for your research, or the more you • What is the overall quality of the
propose research to consider whether are relying on a study for inspiration literature? What have we learned to
these techniques could be adapted for about how to approach your own re- date?
your problem of interest. search, the greater the role it should play • What are the most important prob-
Once you consider literature in this in your literature review. lems and gaps that require addi-
second tier or outer circle, you are likely tional research?
to encounter the problem of too many When there is a lot of literature, it is
sources. The number of potentially rele- not necessary for a review to be compre- These questions are relevant whether
vant publications, especially once you hensive. The literature review should one is producing a stand-alone review
begin considering well-developed areas focus mainly on those parts of the litera- essay, a literature review for a research
of theory, could be vast. Hence, you ture that relate to and help advance your proposal, or a literature review section in
need a way to restrict your focus. It is specific interests; edit out the rest. a finished report such as a thesis or dis-
important not to simply select a few sertation. When one proceeds systemati-
books or articles that you find at random The Bottom Line cally and aims to reach a considered
~for example, whatever happens to be on judgment about the state of knowledge
the library shelf or the first few “hits” A literature review should concisely on a given subject, the resulting literature
returned by a Google search! and make summarize from a set of relevant sources review can itself make a useful contribu-
them the basis for your review, because the collective conclusions most pertinent tion to knowledge.

Notes
* In drafting this overview, I have incorpo- researchers who want to supplement the informa- 4. Paul Pitman first suggested to me this
rated some points made by Paul Pitman in a tion in this essay conduct their own search for phrasing and the reasoning behind it, for which I
lecture delivered to students at the Naval Post- web pages on doing a literature review. thank him.
graduate School. I have also incorporated some 2. The one exception I have found is John- 5. For example, many political science mate-
suggestions contained in a handout prepared by son and Reynolds ~2004, ch. 5!. rials of this sort are available through Political
John Odell for students in the School of Inter- 3. For an introduction to Bayesian statistics, Research Online ~PROL!; this includes papers
national Relations at the University of Southern see Wonnacott and Wonnacott ~1985, 75–79 and presented at Annual Meetings of the American
California. 515–75!. For a discussion of the relevance of Political Science Association ~APSA!. See
1. In Internet searches, however, I have found Bayesian reasoning in qualitative research, see www.politicalscience.org.
several good items. Given the mutability of the McKeown ~1999, 179–83!. 6. This is the first suggestion on a short
Internet, rather than list URLs here, I suggest that handout created by John Odell. It’s a good place

PSOnline www.apsanet.org 131


to start, so I have followed his lead and included Williams ~2003!, “An Appendix on Finding mation and Learning, at www.il.proquest.com0
it first in my own list of pointers. Sources.” umi0dissertations ~accessed Jan. 24, 2005!.
7. For a list of handbooks and encyclopedias 8. The electronic database of UMI disserta- 9. For a discussion of selection bias, see
in many fields of study, see Booth, Colomb, and tions and theses is now part of ProQuest Infor- King, Keohane, and Verba ~1994, 128–39!.

References
Booth, Wayne C., Gregory G. Colomb, and Jo- Galvan, Jose L. 2005. Writing Literature Re- McKeown, Timothy J. 1999. “Case Studies and
seph M. Williams. 2003. The Craft of Re- views, 3rd ed. Glendale, CA: Pyrczak the Statistical Worldview.” International Or-
search, 2nd ed. Chicago: University of Publishing. ganization 53: 161–90.
Chicago Press. Johnson, Janet Buttolph, and H. T. Reynolds. Pan, M. Ling. 2004. Preparing Literature Re-
Cresswell, John W. 2003. Research Design: 2004. Political Science Research Methods, views, 2nd ed. Glendale, CA: Pyrczak
Qualitative, Quantitative, and Mixed Meth- 5th ed. Washington, D.C.: CQ Publishing.
ods Approaches, 2nd ed. Thousand Oaks, Press. Patten, Mildred L. 2005. Understanding Research
CA: SAGE Publications. King, Gary, Robert O. Keohane, and Sidney Methods: An Overview of the Essentials, 5th
Fink, Arlene. 2005. Conducting Research Litera- Verba. 1994. Designing Social Inquiry: ed. Glendale, CA: Pyrczak Publishing.
ture Reviews: From the Internet to Paper, Scientific Inference in Qualitative Wonnacott, Ronald J., and Thomas H. Wonna-
2nd ed. Thousand Oaks, CA: SAGE Research. Princeton, NJ: Princeton Univer- cott. 1985. Introductory Statistics, 4th ed.
Publications. sity Press. New York: John Wiley & Sons.

132 PS January 2006