muse.jhu.edu Open in urlscan Pro
128.220.160.198  Public Scan

Form analysis
3 forms found in the DOM

POST /account/set_attribute_no_ajax/cookie_acknowledgement/1

<form method="post" action="/account/set_attribute_no_ajax/cookie_acknowledgement/1">
  <button type="submit" class="btn_accept" id="accept_cookie_msg">Accept</button>
</form>

POST /search/

<form class="js_off" method="post" action="/search/" style="display: none;">
  <label for="nojs_search_input_header" class="hidden">Search:</label>
  <input name="no_js_header_query" id="nojs_search_input_header">
  <input type="hidden" name="action" value="search">
  <input type="hidden" name="t" value="header">
  <button id="search_button">
    <img src="/images/search_blue.png" alt="search icon">
  </button>
</form>

POST /search/

<form method="post" action="/search/" aria-label="search within this book text input" title="search within this book text input">
  <input type="text" name="no_js_header_query" aria-label="search within this book text input" placeholder="Search Within Book">
  <input type="hidden" name="action" value="search">
  <input type="hidden" name="t" value="search_book_header">
  <input type="hidden" name="search_within_book_id" id="search_within_book_id" value="113313">
  <a id="search_within_book_button">

									<input type="image" alt="a magnifying glass search icon" src="/images/search_blue.png" aria-label="submit search">
								  
								</a>
</form>

Text Content

This website uses cookies to ensure you get the best experience on our website.
Without cookies your experience may not be seamless.

Accept
Accept
[Skip to main content]
Institutional Login
LOG IN
Accessibility
Browse
OR
Search:
Search:
menu
Advanced Search Browse
MyMUSE Account
Log In / Sign Up Change My Account User Settings Access via Institution MyMUSE
Library Search History View History Purchase History MyMUSE Alerts Individual
Subscriptions
Contact Support


INTERNATIONAL ORGANIZATIONS AND RESEARCH METHODS: AN INTRODUCTION



 * Footnotes
 * Chapter
 * University of Michigan Press
 *    
      
      
    * Save Save
      
      
      
      
      
      
      

 * View Citation
 * Additional Information




FOOTNOTES

1.  See for instance the roundtable organized at ISA’s annual conference in 2019
on “Knowledge Without Limits? On Data, Archival Access, Copyright, and Global
Commons in International Relations Research.”

2.  For the detailed program, see the event website:
http://www.unige.ch/run2018/. 

3.  Similar endeavors were last carried out in the 1970s, see Alger (1970) and
Dixon (1977).

4.  The views expressed herein are those of the author(s) and do not necessarily
reflect the views of the United Nations.

5.  http://webtv.un.org/.  (accessed July 31, 2020).

6.  https://enb.iisd.org/.  (accessed July 31, 2020).

7.  Most recently, see for example Pouliot (2016) and Yi-Chong and Weller
(2018).

8.  The tendency of the respondent to provide answers that appear as socially
accepted even if they do not fit with their own opinions.

9.  The difficulty to recall past experiences.

10.  PHAP is a member-based network of self-declared humanitarian professionals
allegedly made up of 10 percent of the global population of aid workers, open to
individuals with at least two years of “relevant experience” in the humanitarian
sector.

11.  Sculpted by Magnús Tómasson; photo by Varfolomeev via Creative Commons.

12.  https://archives.ungeneva.org/lontad.  (accessed February 9, 2023).

13.  https://www.histecon.magd.cam.ac.uk/unhist/index.html.  (accessed June 7,
2022).

14.  The ICRC audiovisual archives. https://avarchives.icrc.org/.  (accessed
October 2, 2020).

15.  The IFRC historic film collection is on Youtube:
https://www.youtube.com/channel/UCd2bE77hqagP0uP2JxfSTdA.  (accessed October 2,
2020).

https://www.who.int/archives/fonds_collections/partners/en/ (accessed October 2,
2020).

17.  The UNHCR visual media center can be found at:
https://media.unhcr.org/C.aspx?VP3=CMS3&VF=Home.  (accessed October 2, 2020).

18.  In 2015, the ICRC put all its audiovisual collection online:
https://avarchives.icrc.org/.  (accessed October 2, 2020).

19.  Most of the time, you find a contact email on the archive resources or
library webpage of the organization.

20.  For example, in 1948 UNICEF partnered for the first time with David “Chim”
Seymour from the newly founded Magnum agency, to attract international attention
and funds for war orphans by using one of the best names in war photography.

21.  See for example the photograph consent form from UNAIDS:
https://www.unodc.org/pdf/india/news_and_events/photographer_guidelines.pdf. 
(accessed October 2, 2020).

22.  See a brief given to research participants by Vince and Warren (Vince and
Warren 2012: 295).

23.  For example, the Indian State of Arunachal Pradesh is isolated from the
rest of the country by a continuous line, the same as for international borders,
because of Chinese claims of this area.

24.  UN system staff members in discussion with the author, July 2018.

25.  For comments and suggestions on this chapter I am grateful to the editors,
the participants at the Geneva Workshop that led to this volume, and
participants at events at Cardiff University and the National University of
Singapore where the core ideas were presented. I thank Humphrey Asamoah for
editorial assistance.

26.  https://uia.org/yearbook. 

27.  https://stats.oecd.org/Index.aspx?DataSetCode=MULTISYSTEM. 

28.  https://www.unsceb.org/content/un-system-financial-statistics. 

29.  http://d-portal.org/ctrack.html#view=search. 

30.  https://www.iatiregistry.org/publisher. 

31.  For the UN system see Dag Hammarskjöld Foundation and UN MPTF Office (2018,
chapter 3).

32.  A good introduction is: (Poole 2005).

33.  For code and examples, see:
https://github.com/evoeten/United-Nations-General-Assembly-Votes-and-Ideal-Points. 

34.  The data used is available on UNHCR data platforms:
http://popstats.unhcr.org/en/overview. 

35.  Formally, N new refugees = N refugees — N refugees in the previous year if
≥ 0, otherwise 0.

36.  High rates of unreported country of origin for UNHCR refugee data is also a
major concern. For a discussion and approaches to overcome this problem, see
Marbach 2018.

37.  The index (polyarchy) measures electoral democracy on a scale from 0
(authoritarian) to 1 (full democracy). We add a squared term to model a possibly
nonlinear relation.

38.  Formally, it measures human rights violations by agents of the state on a
five-point scale.

39.  High-intensity conflicts cause at least a one thousand battle deaths in any
given year. Conversely, low-intensity conflicts are periods during which
violence causes between twenty-five and one thousand battle deaths.

40.  The fixed-effects control for unobservable factors, which may have affected
displacement patterns and are correlated with the independent variables.

41.  While useful in helping to read statistical tables, we would advise
researchers to cautiously approach p-values. Their use has been the target of
criticisms by statisticians (American Statistical Association 2016).

42.  Statistics hinge crucially on key assumptions. It is beyond the scope of
this chapter to discuss these. Interested readers should consult introductory
level statistical literature, such as Agresti (2017).

43.  https://www.unsystem.org/content/un-system-human-resources-statistics. 

44.  Many tutorials exist, but most of them are in French (see for example
Perdoncin 2020). 

45.  Several other metrics are produced such as the cos2—that gives the quality
of the representation of a modality on the plane—and the v-test—a measure of
association between variables.

46.  The number of axes that must be kept in the analysis is both a statistical
and a scientific question. First axes explain more variance than the last ones.
Most of the time, scholars keep in the analysis between two and five axes, the
limit being the axis after which variance explained significatively decreases.

47.  Helpful comments by Gerald Schneider and the editors of this volume are
gratefully acknowledged.

48.  It is useful to note that in IR scholarship there is considerable confusion
about methods and theoretical approaches (or, god forbid, paradigms; see, e.g.,
chapter headings in Reus-Smit and Snidal 2008).

49.  Note also, however, that some national, mostly European, literatures have
dissociated themselves from international trends.

50.  The importance of voting in systematic quantitative studies of IOs is
obviously reflective of the very early quantitative work on voting in
parliaments initiated by among many others Rice (1928).

51.  Thus Brissaud’s (chapter 18—Multiple Correspondence Analysis) reference to
hypotheses will strike many readers as odd, as he is quite explicit about the
fact that MCA proceeds inductively.

52.  We make several key assumptions relating to research design. First, we
assume that the assembled texts are justified as being (limited) representations
of an organization. Second, we assume that the object of interest—such as an
issue, actor, or process—has some observable manifestation within the corpora.
Third, we assume that observable variation in the texts signals something
important for our research question, whether that is a shift in attention or
salience. For all of this, researchers must clearly justify their choices and
theoretical stances.

53.  Different ways of calculating keyness potentially produce different
candidate terms (see Gabrielatos 2018).

54.  See https://labordoc.ilo.org/discovery/search?vid=41ILO_INST:41ILO_V2.  A
small number of reports were not machine-readable or published during this
period.

55.  See Allen and Blinder (2018) for an example of applying similar
computational techniques in a political communication setting.

56.  The views expressed herein are those of the author and do not necessarily
reflect the views of the United Nations.

57.  https://sustainabledevelopment.un.org/mgos.  (accessed March 31, 2021).

58.  See, for example, IO BIO—Biographical Dictionary of Secretaries-General of
International Organizations project, www.ru.nl/fm/iobio. 

59.  Search handbooks such as Baker and Sibonile Ellece (2011) to identify
relevant DA tools.

60.  King, Keohane, and Verba 1994 refer to such studies as qualitative but then
go on to argue that the rules of inference that inform such studies should be
the same as those in quantitative inferential research.

61.  I use neopositivist here following the definition of the term in Jackson
2011, with its focus on hypothesis testing.

62.  Epistemology referring here to general ideas about how we should go about
studying and understanding the world, as distinct from methodology, which is a
strategy to address a specific question. Epistemology in this sense is roughly
equivalent to what Jackson 2011 refers to as philosophical ontology.

63.  Theory understood here in a broad and inclusive sense, to forestall
argument about what theory really is.

. Research is often—if not always—a story of informal relationships, friendship,
luck, and (seizing) opportunities. This book project is no exception. The first
seed of a book on research methods and IOs dates back to the spring of 2016 when
Fanny Badache failed to find the right methodological handbook while starting
her PhD dissertation. It was soon taken on as a collective adventure among four
women (Fanny Badache, Emilie Dairon, Leah R. Kimber, and Lucile
Maertens)—colleagues and friends—passionate by the research on international
organizations (IOs) #IOGeeks. Scientifically, the idea of this book results from
a two-fold observation. On the one hand, many IO scholars have recently called
for a renewal in the way IOs are approached, stressing the need for innovative
tools to capture the intergovernmental and transnational world. On the other
hand, the same scholars have highlighted the difficulty to conduct research on
IOs, due to the need for “discretion” and access barriers.1 To contribute to
these discussions, we felt compelled to bridge scientific networks (mainly
divided by linguistic differences/barriers) and to gather scholars working on
IOs but yet embedded in different academic contexts.

. 

(2) Convenience sampling: Use readily available mailing lists, such as the
Professionals in Humanitarian Assistance and ProtectionPage 74 → (PHAP)10
mailing list, or social media groups, such as the “International Humanitarian
and Development Professionals” LinkedIn group.



. Page 96 →Personifying IOs. Close your eyes and imagine the United Nations.
What comes to mind? Perhaps an azimuthal projection of the globe cradled in
olive tree branches, long rows of state flags, or the newly renovated Security
Council chamber for the UN connoisseur. But what about the UN workforce? Like
the unidentifiable civil servant holding a briefcase portrayed in Magnús
Tómasson’s sculpture Unknown Bureaucrat,11 the IO workforce is, for the most
part, faceless (Reinalda, Kille, and Eisenberg 2018). By asking interviewees
about personal matters, such as their youth or what motivated them to pursue an
international career, biographic interviews give the means to personify and
deimpersonalize IOs (Emmerij et al. 2005; Yi-Chong and Weller 2018).

. Some archives are taking steps to remedy these problems by digitizing their
collections. The UN Library in Geneva, for example, now offers online access to
the entire League of Nations archival collection.12 The International
Telecommunication Union (ITU) is also making parts of the ITU’s documentation
available online. Several other archives have digitized parts of their
collections, and these are mostly free to download through their online search
catalogues. For a useful overview of relevant online archives for IO research,
see the website of the United Nations History Project.13

. Some archives are taking steps to remedy these problems by digitizing their
collections. The UN Library in Geneva, for example, now offers online access to
the entire League of Nations archival collection.12 The International
Telecommunication Union (ITU) is also making parts of the ITU’s documentation
available online. Several other archives have digitized parts of their
collections, and these are mostly free to download through their online search
catalogues. For a useful overview of relevant online archives for IO research,
see the website of the United Nations History Project.13

. Since the 1860s–1870s, drawings, paintings, posters, daguerreotypes,
postcards, stamps, pictures, films but also leaflets and periodicals are
preserved by the International Committee of the Red Cross (ICRC)14 or the
International Federation of the Red Crescent (IFRC),15 and also by the United
Nations where each agency has its own visual archives (such as the United
Nations Refugee Agency or the World Health Organization). Exploration of visual
archives can enrich or fill gaps, as well as open new lines of enquiry over the
imperialist and colonialist ideologies of the late nineteenth century, or the
settings of multilateral diplomacy in the interwar period of North-South
relations after 1945. For example, a historical investigation over the poster
archives of the International Red Cross and Red Crescent Museum—the largest
humanitarian collection in the world—has questioned the visual politics and the
reappropriation of humanitarian principles throughout the Red Cross movement
since the First World War (Gorin 2019).

. Since the 1860s–1870s, drawings, paintings, posters, daguerreotypes,
postcards, stamps, pictures, films but also leaflets and periodicals are
preserved by the International Committee of the Red Cross (ICRC)14 or the
International Federation of the Red Crescent (IFRC),15 and also by the United
Nations where each agency has its own visual archives (such as the United
Nations Refugee Agency or the World Health Organization). Exploration of visual
archives can enrich or fill gaps, as well as open new lines of enquiry over the
imperialist and colonialist ideologies of the late nineteenth century, or the
settings of multilateral diplomacy in the interwar period of North-South
relations after 1945. For example, a historical investigation over the poster
archives of the International Red Cross and Red Crescent Museum—the largest
humanitarian collection in the world—has questioned the visual politics and the
reappropriation of humanitarian principles throughout the Red Cross movement
since the First World War (Gorin 2019).

. Most of the time, visual artifacts are preserved in separate archival units,
collections, or buildings, even in the same organization, depending on the
visual or textual nature of the source.16 The constellation of archival sites
thus represents multiple challenges for the researcher. It is greatly advised to
contact the archivist in charge of the visual archives in the organization (see
chapter 8—Archives). Some archives are digitized online for free, but they Page
133 →do not include all visual artifacts (e.g., memorabilia). An on-site visit
to the archives is necessary for researchers interested in two aspects: first, a
focus on the materiality of visual objects, such as the technique or the
constraints related to viewing devices, will allow researchers to understand the
immersive experience of spectators from previous decades; second, the
consultation of written material accompanying images is mandatory (e.g.,
reports, personal correspondence, field notes) to have contextual information
about the production/dissemination process (see chapter 9—Visual Methods).
Visual archives preserved in the same location offer homogeneous corpuses that
give meaningful inputs about the use and reuse of an image, sometimes for
multiple purposes and through different visual formats. It is therefore useful
to explore the political and cultural specificities of images and their
variations through space and time. However, it is not always possible to trace
the history of their creation. Many images before 1945 are preserved without any
captions, nor location, date, or name of the image-maker. Finally, accessibility
is also an issue: images of the early twentieth century are accessible because
they fall into public domain. However, there might be a limitation period of
approximately twenty to forty years on contemporary images for confidentiality
reasons.

. If you plan to analyze images, the first step aims at identifying
institutional archives, either physical or digital, by establishing a list of
major IOs related to your topic (see also box l—Visual Archives). To find out
visual databases, you have three options: whether the organization has a library
or documentation center, archive units, or a media center. Many visual databases
are either unrecognized or underused by researchers interested in international
relations—such as the UNHCR17 or the ICRC.18 Many of them are open to the
public, but if not you should seek whether there is an exception for scholars.
In many cases, a visit to the physical archives or library in the headquarters
could be useful. Although using online research engines or databases is helpful,
sometimes they do not include everything. Therefore it is also rewarding to
contact librarians working in these organizations,19 so they can help you
identify specific collections and get familiar with the databases. The secondary
literature can also be useful (especially in history, ethnography, political
sciences, or international relations) to help locate visual materials previously
used or mentioned by researchers. Make sure you find visual databases with
sufficient information about image producers (e.g., personal biographies,
mandate framework with the IO),20 the context of images, their dissemination,
and possibly the targeted public.

. If you plan to analyze images, the first step aims at identifying
institutional archives, either physical or digital, by establishing a list of
major IOs related to your topic (see also box l—Visual Archives). To find out
visual databases, you have three options: whether the organization has a library
or documentation center, archive units, or a media center. Many visual databases
are either unrecognized or underused by researchers interested in international
relations—such as the UNHCR17 or the ICRC.18 Many of them are open to the
public, but if not you should seek whether there is an exception for scholars.
In many cases, a visit to the physical archives or library in the headquarters
could be useful. Although using online research engines or databases is helpful,
sometimes they do not include everything. Therefore it is also rewarding to
contact librarians working in these organizations,19 so they can help you
identify specific collections and get familiar with the databases. The secondary
literature can also be useful (especially in history, ethnography, political
sciences, or international relations) to help locate visual materials previously
used or mentioned by researchers. Make sure you find visual databases with
sufficient information about image producers (e.g., personal biographies,
mandate framework with the IO),20 the context of images, their dissemination,
and possibly the targeted public.

. If you plan to analyze images, the first step aims at identifying
institutional archives, either physical or digital, by establishing a list of
major IOs related to your topic (see also box l—Visual Archives). To find out
visual databases, you have three options: whether the organization has a library
or documentation center, archive units, or a media center. Many visual databases
are either unrecognized or underused by researchers interested in international
relations—such as the UNHCR17 or the ICRC.18 Many of them are open to the
public, but if not you should seek whether there is an exception for scholars.
In many cases, a visit to the physical archives or library in the headquarters
could be useful. Although using online research engines or databases is helpful,
sometimes they do not include everything. Therefore it is also rewarding to
contact librarians working in these organizations,19 so they can help you
identify specific collections and get familiar with the databases. The secondary
literature can also be useful (especially in history, ethnography, political
sciences, or international relations) to help locate visual materials previously
used or mentioned by researchers. Make sure you find visual databases with
sufficient information about image producers (e.g., personal biographies,
mandate framework with the IO),20 the context of images, their dissemination,
and possibly the targeted public.

. As a result, in the fall of 2017 a call for papers for an international
workshop on the theme “Researching the United Nations: Rethinking Methods of
Investigation” (RUN2018) was widely distributed. We successfully received more
than 110 abstracts leading to a two-day workshop organized at the University of
Geneva in June 2018 with more than seventy researchers from various countries
and academic communities. The workshop consisted Page xx →of fourteen panels and
two plenary sessions.2 A few months later, the project for the present book was
launched.

. If you plan to analyze images, the first step aims at identifying
institutional archives, either physical or digital, by establishing a list of
major IOs related to your topic (see also box l—Visual Archives). To find out
visual databases, you have three options: whether the organization has a library
or documentation center, archive units, or a media center. Many visual databases
are either unrecognized or underused by researchers interested in international
relations—such as the UNHCR17 or the ICRC.18 Many of them are open to the
public, but if not you should seek whether there is an exception for scholars.
In many cases, a visit to the physical archives or library in the headquarters
could be useful. Although using online research engines or databases is helpful,
sometimes they do not include everything. Therefore it is also rewarding to
contact librarians working in these organizations,19 so they can help you
identify specific collections and get familiar with the databases. The secondary
literature can also be useful (especially in history, ethnography, political
sciences, or international relations) to help locate visual materials previously
used or mentioned by researchers. Make sure you find visual databases with
sufficient information about image producers (e.g., personal biographies,
mandate framework with the IO),20 the context of images, their dissemination,
and possibly the targeted public.

. If you plan to generate images, you have to decide whether you want to produce
images yourself or if you need to include research participants within the
organization. Many layers of organizational settings can be directly observed
and visually recorded: employees, buildings, meetings, interactions, processes.
Try to remain unobtrusive; this can happen with photographic devices or Page 138
→videotaped recordings, so you might opt for drawings, diagrams, or graphs
instead (Meyer 1991). Conversely, you can include participants to make images of
places that would otherwise remain inaccessible to the researcher. Thus Samantha
Warren asked members of an IT firm to picture their desks to explore the
aesthetics of work environments (Warren 2002). This option will include field
preparation with the IOs and instructions to participants. This is a very
important aspect: you need to have a mutual discussion over the set-up and
execution of the study. Some legal and ethical issues have to be agreed upon;
for example, you need to know whether subjects will be aware they are being
recorded and whether they agree (informed consent). Many IOs have working
agreements already established with photographers that might be useful and
readjusted by researchers, whether to obtain consent or to pay attention to
ethical issues when using or producing images.21

. Page 139 →If visual data were generated to include the “voices” of the
stakeholders, the researcher has to question the encoding-decoding process. This
means to plan qualitative interviews with the research participants, so they can
explain their views, their choices, what they take for granted. It allows the
researcher to avoid assumptions, bias, or question elements that might not be
obvious or relevant to him/her.22

. In the UN context, the Geospatial Information Section, responsible for
clearance process of any UN map, sets standards (graphic and toponymical) about
accepted international borders. UN maps thus reflect power dynamics and issues
among its members through footnotes on territorial disputes, dotted lines for
undetermined boundaries, or even continuous lines inside sovereign states.23
Otherwise, the notorious reports of the Intergovernmental Panel on Climate
Change skillfully maps a borderless Earth in order to alert the public about the
global dimension of the planet’s warming, while at the same time gridding the
world according to countries when states have to take actions. Maps mostly serve
as medium for adopted models on future possible climate scenarios.

. The second question explores the factors that shape branding. Questions on why
an organization chooses its name or its logo help explain how an IO’s
characteristics and context influence branding manifestations. One such factor
may be funding. Some agencies in the UN system, such as UNICEF, are considered
to receive more funding from private donors due to their perceived status as
entities not closely associated with the UN.24 In keeping with this argument,
maintaining its current name would be more beneficial for UNICEF, rather than
entertaining a name change to “UN Children.”

. 


WHAT?25



. Various ways exist for measuring IO budgets, revenue, and expenditures. For
more recent periods, myriad datasets exist with varying levels of detail,
accuracy, and coverage. Common, but not necessarily reliable and valid, sources
are the Yearbook of International Organizations26 (includes annual budget
figures but without clear-cut budget definitions or expenditure data); the
OECD-DAC databases27 (not always comprehensive and limited time coverage); the
financial statistics provided by UNSCEB for the UN system28 (with consistency
issues going back in time); and individual IOs’ websites with Page 189
→downloadable budget sheets (not necessarily consistent over time). Recently,
the International Aid Transparency Initiative (IATI) has started providing
additional data, especially at the project level29 but also by agency.30

. Various ways exist for measuring IO budgets, revenue, and expenditures. For
more recent periods, myriad datasets exist with varying levels of detail,
accuracy, and coverage. Common, but not necessarily reliable and valid, sources
are the Yearbook of International Organizations26 (includes annual budget
figures but without clear-cut budget definitions or expenditure data); the
OECD-DAC databases27 (not always comprehensive and limited time coverage); the
financial statistics provided by UNSCEB for the UN system28 (with consistency
issues going back in time); and individual IOs’ websites with Page 189
→downloadable budget sheets (not necessarily consistent over time). Recently,
the International Aid Transparency Initiative (IATI) has started providing
additional data, especially at the project level29 but also by agency.30

. Various ways exist for measuring IO budgets, revenue, and expenditures. For
more recent periods, myriad datasets exist with varying levels of detail,
accuracy, and coverage. Common, but not necessarily reliable and valid, sources
are the Yearbook of International Organizations26 (includes annual budget
figures but without clear-cut budget definitions or expenditure data); the
OECD-DAC databases27 (not always comprehensive and limited time coverage); the
financial statistics provided by UNSCEB for the UN system28 (with consistency
issues going back in time); and individual IOs’ websites with Page 189
→downloadable budget sheets (not necessarily consistent over time). Recently,
the International Aid Transparency Initiative (IATI) has started providing
additional data, especially at the project level29 but also by agency.30

. Various ways exist for measuring IO budgets, revenue, and expenditures. For
more recent periods, myriad datasets exist with varying levels of detail,
accuracy, and coverage. Common, but not necessarily reliable and valid, sources
are the Yearbook of International Organizations26 (includes annual budget
figures but without clear-cut budget definitions or expenditure data); the
OECD-DAC databases27 (not always comprehensive and limited time coverage); the
financial statistics provided by UNSCEB for the UN system28 (with consistency
issues going back in time); and individual IOs’ websites with Page 189
→downloadable budget sheets (not necessarily consistent over time). Recently,
the International Aid Transparency Initiative (IATI) has started providing
additional data, especially at the project level29 but also by agency.30

. To what extent have scholars discussed research methods in their studies of
IOs? How do they present their methodology? Can we observe an evolution over the
past seventy-five years? To answer these questions, we conducted a longitudinal
study in which we analyzed a corpus of research articles on IOs from 1945, the
year of the creation of the United Nations, to 2020.3 The corpus is based on
articles drawn from the seven most used and referenced peer-reviewed journals
according to the contributors to this book. Since 1950 and following a five-year
mark, we retrieved the articles addressing IOs and/or multilateralism in either
the title or the abstract (see table 1 below).

. Various ways exist for measuring IO budgets, revenue, and expenditures. For
more recent periods, myriad datasets exist with varying levels of detail,
accuracy, and coverage. Common, but not necessarily reliable and valid, sources
are the Yearbook of International Organizations26 (includes annual budget
figures but without clear-cut budget definitions or expenditure data); the
OECD-DAC databases27 (not always comprehensive and limited time coverage); the
financial statistics provided by UNSCEB for the UN system28 (with consistency
issues going back in time); and individual IOs’ websites with Page 189
→downloadable budget sheets (not necessarily consistent over time). Recently,
the International Aid Transparency Initiative (IATI) has started providing
additional data, especially at the project level29 but also by agency.30

. The central challenge, however, regards the availability, quality, and
comparability of budget data.31 Constructing a dataset with reliable financial
data going past the last decade or so can be fraught with difficulties. While
websites of IOs have improved in recent years, with some even making
downloadable financial datasets available, finding official figures can be hard
unless there exists a dedicated and up-to-date IO budget website.

. Empirical spatial models simultaneously estimate ideal points and roll-call
parameters from observed vote choices. There is a vast literature explaining
these models, even though most of it focuses on legislatures and courts rather
than IOs.32 There are also numerous options to estimate these models. This
ranges from traditional ready-made models that can be estimated in R or Stata,
such as W-NOMINATE (Poole et al. 2008), to customizable Bayesian frameworks
(Grant et al. 2017).

. Third, analysts have to decide on dimensionality. In the context of the UN
General Assembly, Michael Bailey and I examined whether there is a stable,
important, and interpretable second dimension (Bailey and Voeten 2018). We found
that from the mid-1960s to the mid-1980s, North–South conflict constitutes a
stable second dimension. In the periods before and after, the second dimension
neither is stable nor easily interpretable, though it is sometimes important. In
most applications, one-dimensional estimates have conceptual advantages with
minimal losses in explanatory value. Yet analysts who care about particular
issues may well consider estimating ideal point models on subsamples of the
data.33 For example, voting behavior on nuclear issues often poorly fits the
first dimension.

. Over the last two decades, international organizations (IOs) have become major
data providers. This data is instrumental in supporting scientific research and
shaping public policies. In this chapter, we discuss how statistical analyses of
IO data can help answer important research questions, such as “what determines
refugee flows?” We illustrate our argument by using data from the United Nations
High Commissioner for Refugees (UNHCR).34

. As a first step, we encourage researchers to review the documentation to
understand the definitions employed by IOs, the format of the data and possible
limitations. For a concrete example, UNHCR provides information on refugee
stocks in a dyadic format (pairs of countries of origin—asylum). This has two
implications. First, for applications (as our own) in which researchers study
the determinants of refugee flows, it is useful to aggregate the data by country
of origin, resulting thus in a monadic format. Second, while the data measures
refugee stocks in a given period, we are interested in refugee flows. Hence we
generate a variable measuring the number of new refugees by subtracting the
number of refugees in the previous year from the total for the next year.
Negative refugee counts are censored at zero.35,36

. As a first step, we encourage researchers to review the documentation to
understand the definitions employed by IOs, the format of the data and possible
limitations. For a concrete example, UNHCR provides information on refugee
stocks in a dyadic format (pairs of countries of origin—asylum). This has two
implications. First, for applications (as our own) in which researchers study
the determinants of refugee flows, it is useful to aggregate the data by country
of origin, resulting thus in a monadic format. Second, while the data measures
refugee stocks in a given period, we are interested in refugee flows. Hence we
generate a variable measuring the number of new refugees by subtracting the
number of refugees in the previous year from the total for the next year.
Negative refugee counts are censored at zero.35,36

. In a third step, we use the available data from UNHCR to determine to what
extent authoritarian regimes, human rights violations, and armed conflicts cause
displacement using linear regression. As the name indicates, we assume a (log)
linear relationship between these variables and displacement. Data for these
variables comes from the vdem dataset (Coppedge et al. 2019),37 the Political
Terror Scale (Gibney et al. 2019),38 and the UCDP Armed Conflict Dataset (v
19.1) (Gleditsch et al. 2002; Pettersson, Högbladh, and Öberg 2019). In addition
to low and high intensity, we distinguish between interstate and internal
conflicts.39 Finally, we control for GDP per capita and population (World Bank
2019) and add country and year fixed effects.40

. In a third step, we use the available data from UNHCR to determine to what
extent authoritarian regimes, human rights violations, and armed conflicts cause
displacement using linear regression. As the name indicates, we assume a (log)
linear relationship between these variables and displacement. Data for these
variables comes from the vdem dataset (Coppedge et al. 2019),37 the Political
Terror Scale (Gibney et al. 2019),38 and the UCDP Armed Conflict Dataset (v
19.1) (Gleditsch et al. 2002; Pettersson, Högbladh, and Öberg 2019). In addition
to low and high intensity, we distinguish between interstate and internal
conflicts.39 Finally, we control for GDP per capita and population (World Bank
2019) and add country and year fixed effects.40

. In a third step, we use the available data from UNHCR to determine to what
extent authoritarian regimes, human rights violations, and armed conflicts cause
displacement using linear regression. As the name indicates, we assume a (log)
linear relationship between these variables and displacement. Data for these
variables comes from the vdem dataset (Coppedge et al. 2019),37 the Political
Terror Scale (Gibney et al. 2019),38 and the UCDP Armed Conflict Dataset (v
19.1) (Gleditsch et al. 2002; Pettersson, Högbladh, and Öberg 2019). In addition
to low and high intensity, we distinguish between interstate and internal
conflicts.39 Finally, we control for GDP per capita and population (World Bank
2019) and add country and year fixed effects.40

. Kari De Pryck and Svenja Rauch4

. In a third step, we use the available data from UNHCR to determine to what
extent authoritarian regimes, human rights violations, and armed conflicts cause
displacement using linear regression. As the name indicates, we assume a (log)
linear relationship between these variables and displacement. Data for these
variables comes from the vdem dataset (Coppedge et al. 2019),37 the Political
Terror Scale (Gibney et al. 2019),38 and the UCDP Armed Conflict Dataset (v
19.1) (Gleditsch et al. 2002; Pettersson, Högbladh, and Öberg 2019). In addition
to low and high intensity, we distinguish between interstate and internal
conflicts.39 Finally, we control for GDP per capita and population (World Bank
2019) and add country and year fixed effects.40

. Table 6 presents the results of the regression. The first estimate for each
variable is the coefficient, which can be interpreted as our “best guess” for
the direction and magnitude of the relation between a predictor and the
dependent variable. The number in parentheses below is the standard error, which
is an estimate of uncertainty. The larger the standard error relative to the
coefficient, the more uncertain the association. Asterisks indicate whether the
T value (the ratio between the coefficient and the standard error) has crossed a
p-value threshold.41

. Conducting research using IO data raises specific challenges. We discuss three
aspects: definitions, measurement, and the unit of analysis. Researchers should
be alert to these concerns and carefully assess the validity and reliability of
their data.42 While we illustrate our discussion with UNHCR Page 218 →data on
refugee flows, readers should be mindful that concerns pertaining to data
produced by IOs are frequently idiosyncratic to one’s own specific line of
inquiry. Readers are therefore invited to carefully appraise IO data in their
own field of investigation.

. The second available source includes the statistics gathered by the Chief
Executives Board for Coordination.43 They provide human resource data for all UN
organizations on age, staff location, nationality, gender, length of Page 222
→service, grade, and category of service. These statistics are graphically
presented on the website and can be downloaded. This database is very useful to
make comparisons between UN organizations. However, since they are presented by
organizations and by characteristics, they cannot be used for cross-sectional
analyses such as studies that address the composition in terms of age and
gender. Another limitation of this set of HR data is that they are limited to
staff with fixed-term contracts of one year of more.

. Once the dataset is constructed, MCA can be run thanks to different platforms
such as the package FactoMineR (Lê et al. 2008) from the statistical software
R.44 Applied to the database, MCA produces a table that gives, for each modality
of variable, its contribution to the production of axes.45 It also produces
another table that gives the positions of modalities of variables and the
positions of individuals on the graph. Then it plots a two-dimensional graph
that must be improved using other packages such as ggplot2, showing the
modalities of variables on a plane. The researcher must then look at the
percentages of variance explained by every axis. It is a measure of the
proportion of the differences between individuals that is summarized by the
axis. Then the researcher chooses the number of axes that must be retained in
the analysis (often two or three).46 The plots should be read Page 242 →as an
opposition on the west-east axis (the x axis), and on the north-south axis (the
y axis), with each part of the graph describing a subpopulation. It is often
useful to exemplify these subpopulations with the description of the trajectory
of one of the individuals that are positioned here. The following figure
illustrates such results based on my research. Due to space restriction, I have
limited the presentation of variables to those represented in the first two axes
of the MCA, which gather most of the overall information.

. Once the dataset is constructed, MCA can be run thanks to different platforms
such as the package FactoMineR (Lê et al. 2008) from the statistical software
R.44 Applied to the database, MCA produces a table that gives, for each modality
of variable, its contribution to the production of axes.45 It also produces
another table that gives the positions of modalities of variables and the
positions of individuals on the graph. Then it plots a two-dimensional graph
that must be improved using other packages such as ggplot2, showing the
modalities of variables on a plane. The researcher must then look at the
percentages of variance explained by every axis. It is a measure of the
proportion of the differences between individuals that is summarized by the
axis. Then the researcher chooses the number of axes that must be retained in
the analysis (often two or three).46 The plots should be read Page 242 →as an
opposition on the west-east axis (the x axis), and on the north-south axis (the
y axis), with each part of the graph describing a subpopulation. It is often
useful to exemplify these subpopulations with the description of the trajectory
of one of the individuals that are positioned here. The following figure
illustrates such results based on my research. Due to space restriction, I have
limited the presentation of variables to those represented in the first two axes
of the MCA, which gather most of the overall information.

. Once the dataset is constructed, MCA can be run thanks to different platforms
such as the package FactoMineR (Lê et al. 2008) from the statistical software
R.44 Applied to the database, MCA produces a table that gives, for each modality
of variable, its contribution to the production of axes.45 It also produces
another table that gives the positions of modalities of variables and the
positions of individuals on the graph. Then it plots a two-dimensional graph
that must be improved using other packages such as ggplot2, showing the
modalities of variables on a plane. The researcher must then look at the
percentages of variance explained by every axis. It is a measure of the
proportion of the differences between individuals that is summarized by the
axis. Then the researcher chooses the number of axes that must be retained in
the analysis (often two or three).46 The plots should be read Page 242 →as an
opposition on the west-east axis (the x axis), and on the north-south axis (the
y axis), with each part of the graph describing a subpopulation. It is often
useful to exemplify these subpopulations with the description of the trajectory
of one of the individuals that are positioned here. The following figure
illustrates such results based on my research. Due to space restriction, I have
limited the presentation of variables to those represented in the first two axes
of the MCA, which gather most of the overall information.

. 


INTRODUCTION47



. Let me also be clear what I consider to be a method: a method is a set of
procedures that allows us to adjudicate among rival hypotheses (Sprinz and
Wolinsky-Nahmias 2004: 4). So a method is not a technique of data collection
alone (e.g., participant observations, elite surveys, text or network analysis,
etc.) nor is it a theory (see Sprinz and Wolinsky-Nahmias 2004 and Barkin in
this volume, Interlude V—Controversies on Methodological Pluralism, though this
latter text unduly associates methods with epistemological stances).48 Thus if
one considers the use of quantitative methods in the study of IOs, one generally
finds almost always related (though often delayed) patternsPage 256 → as in IR
or political science more broadly. Thus specific quantitative tools were
introduced (or discarded) in political science or IR more generally, before
similar introductions (and eliminations) occurred in research on IOs. While
Singer and Alger’s (1968: 3) quip that “[I]f it can be measured, it must be
trivial” has been made to characterize the critique of quantitative work in IR
more broadly, it certainly also characterized scholarship on IOs critical of
quantitative methods more specifically.

. Thus in some sense the IO literature followed a similar pattern as many other
areas of IR scholarship, from mostly qualitative case studies to systematic
comparisons and finally studies taking inferential problems seriously. This is
reflected in the increasing importance of quantitative work (see Sprinz and
Wolinsky-Nahmias 2004: 4). Nevertheless, quantitative scholars in IR more
generally and IOs (Hafner-Burton, von Stein, and Gartzke 2008) more specifically
still feel the need to defend their approaches.49

. First, because direct observation produces data about activities that usually
take place behind closed doors (and sometimes even in secret settings),
researchers must deal carefully with data collection and disclosure. In this
context, IOs and universities increasingly ask researchers to provide
information on data protection, to ensure that the identity of the individuals
that are included in their studies will not be displayed without their consent.
When investigating multilateral negotiations, researchers must also carefully
treat any information that could impede on the process. A technique that is
commonly used to circumvent these challenges is to complement findings from
direct observation with other data, including interviews, personal accounts
(such as biographies and books), official documents, and meeting records and
webcasts (see part 2—Interviewing, and chapter 20—Interviews and Observations).5
When working on global environmental negotiations, for instance, a particularly
relevant source of information are the reports Page 26 →of the Earth
Negotiations Bulletin, produced by the International Institute for Sustainable
Development (IISD).6 IISD releases daily or weekly summaries of major
multilateral negotiations on the environment and on sustainable development,
providing relatively neutral records on the ongoing deliberations.

. As the review of systematic work on IOs in the 1950s and 1960s by Alger (1970:
444) nicely shows, analyses of voting, mostly in the General Assembly of the UN
(UNGA), dominated (see review offered by Voeten in chapter 14—Voting
Analysis),50 which led this author also to a critical assessment:

. The method we outline derives inspiration from the fields of computational and
corpus linguistics, which seek to identify patterns of language use in sets of
text (called “corpora”) that are systematically assembled for particular
purposes (Taylor and Marchi 2018). Our approach uses these patterns both as
guides to potentially relevant portions of an archive and as forms of knowledge
in their own right that can be used to address questions that Page 286
→researchers ask about international organizations.52 First, if researchers are
already interested in a particular population, concept, or event (e.g.,
refugees, democracy), descriptive patterns can be identified of where those
terms of interest tend to appear more or less frequently as a proportion of all
the text in each document. Then more intensive close reading of those areas
containing more mentions can be undertaken. By contrast, if researchers are
interested in identifying terms that might characterize one portion of the
archive (e.g., which issues appear more often in a certain time period), the
frequencies of the words in one subset of documents can be compared to their
frequencies in another subset. Having found those terms, researchers can
scrutinize them further using qualitative methods.

. We use three types of linguistic analysis arranged in different orders
depending on the research goal. Frequency analysis counts instances of a given
word or group of words and then displays these frequencies either in terms of
the whole corpus or differentiated by subsets of texts (“subcorpora.”) To enable
comparison of frequencies arising from differently sized subcorpora, these
frequencies are normalized into occurrences per thousand or million words
(Taylor and Marchi 2018). Keyword analysis identifies terms that are central to
a corpus or subcorpus by comparing it to another set of texts called a
“reference corpus.” Words more unique to the corpus in question are better
candidates for exploring what the texts might be about.53 Finally, concordance
analysis displays the text around every instance of a word or phrase of
interest,Page 288 → whether this interest is determined by frequency, keyness,
or another measure. This provides a window onto how a given word or phrase is
used in the documents, and a way to gain further qualitative access for either
interpretation or disambiguation (for example, between asylum as used in either
forced migration or mental health settings and environment to refer to political
contexts rather than ecological terms).

. Our working example involves analyzing a corpus of ninety-three annual reports
produced by the International Labour Organization (ILO) between 1919 and 2015.
These documents, made available online through the ILO’s digital archive,
comprise more than eighteen million words.54 Our unit of analysis are individual
reports, meaning that each year is represented by one document. While we did not
divide the dataset into subcorpora (such as by decade), depending on the
research question and types of archive materials, this is an option worth
considering.

. Second, quantitative techniques—even relatively simple ones such as counting
words—do not absolve researchers of either avoiding serious qualitative and
interpretive work or accounting for their own (and possibly unnoticed) biases in
making those interpretations. Moreover, our approach needs to be embedded within
strong designs and theoretical expectations as appropriate for researchers’
disciplinary norms.55 Although computational techniques may provide signposts
toward key parts of the archive, we advise not to underestimate the amount of
time and energy needed to make sense of these patterns and show their wider
significance for a given discipline or audience.

. Svenja Rauch56

. To identify the overall narrative that underpins the 2030 Agenda, I used
discourse analysis (see chapter 11—Discourse Analysis) to study statements
issued by the parties to the negotiations, the various iterations of the
negotiation text, and accompanying documentation provided by UN system entities,
academia, and major groups and other stakeholders.57 Discourse analysis
(WiddowsonPage 312 → 2007), on the one hand, helps discover “how meaning is
negotiated between the members of a discourse community” (Angermuller,
Maingueneau, and Wodak 2014: 3). On the other hand, it is critical to understand
the technical language of resolutions, negotiation documents, and reports to
reveal the delicate political balance struck in the respective texts.

. Direct observation can further prepare the ground for interviews to generate
additional qualitative data. Interview-based research helps elicit the views of
individuals involved in the process in a limited timeframe (see chapter
5—Semistructured Interviews). The semistructured format offers “sufficient
flexibility to accommodate the individual circumstances of each interview, while
also providing a basic structure of questions that allows for comparison” (Lodge
2013: 187). With a set of preestablished questions, researchers may then tailor
them to each interviewee. As interviews with international elites are—most of
the time—“one-shot games” (Lodge 2013: 194) and involve power asymmetries, each
interview needs to be prepared thoroughly. Prosopography (see chapter
26—Prosopography) can be a useful tool in this regard.58

. Second, you empirically assess your writing—e.g., drafts—via discourse
analysis to see whether the implicit dimensions of the discourses you produce
match your compass-discourse. To investigate whether you inadvertently deny the
agency of so-called local actors, you may, for example, look Page 344 →for
subject positionings (e.g., who are the subjects in my text? What are the
adjectives and verbs associated with these subjects?), and processes of
passivization (i.e., the transformation of potentially active sentences into
passive ones leading to the discursive disempowering of the actors being
passivized).59

. First, because direct observation produces data about activities that usually
take place behind closed doors (and sometimes even in secret settings),
researchers must deal carefully with data collection and disclosure. In this
context, IOs and universities increasingly ask researchers to provide
information on data protection, to ensure that the identity of the individuals
that are included in their studies will not be displayed without their consent.
When investigating multilateral negotiations, researchers must also carefully
treat any information that could impede on the process. A technique that is
commonly used to circumvent these challenges is to complement findings from
direct observation with other data, including interviews, personal accounts
(such as biographies and books), official documents, and meeting records and
webcasts (see part 2—Interviewing, and chapter 20—Interviews and Observations).5
When working on global environmental negotiations, for instance, a particularly
relevant source of information are the reports Page 26 →of the Earth
Negotiations Bulletin, produced by the International Institute for Sustainable
Development (IISD).6 IISD releases daily or weekly summaries of major
multilateral negotiations on the environment and on sustainable development,
providing relatively neutral records on the ongoing deliberations.

. However, the distinction between quantitative and qualitative is often drawn
at the level of methodology or research design (see chapter 25—Process Tracing).
At this level, it makes even less sense than at the level of method. In this
usage, “quantitative” is usually meant to describe a research design that uses
statistical modeling to inferentially test formally stated hypotheses.
Qualitative methodologies might then be understood as a default category
covering everything else, or more narrowly as a research design that is
interpretive in intent. A quantitative research design understood in this way is
likely to include a quantitative method narrowly defined. But many research
designs that use statistics would not be included (anything, for example, that
uses descriptive statistics but is not designed as an exercise in inferential
hypothesis testing). Furthermore, it is not clear where a set of comparative
case studies designed to test hypotheses would fit, since the logic of the
research design is inferential rather than interpretive.60

. What people often mean when they refer to a quantitative methodology, in other
words, is neither quantitative per se nor a methodology. It may be better
understood as inferential or neopositivist,61 and as epistemology62 rather than
methodology. It might be contrasted most simply with an interpretive
epistemology, or more robustly with a range of other understandings of what we
are trying to do when we do social science (Jackson 2011). This might sound at
first like a terminological quibble, but confusing our methods with our
methodologies with our epistemologies can, and often does, put unreasonable
limits on our ability to think creatively and productively about using methods
and about building the methodologies that will most effectively address our
research questions.

. What people often mean when they refer to a quantitative methodology, in other
words, is neither quantitative per se nor a methodology. It may be better
understood as inferential or neopositivist,61 and as epistemology62 rather than
methodology. It might be contrasted most simply with an interpretive
epistemology, or more robustly with a range of other understandings of what we
are trying to do when we do social science (Jackson 2011). This might sound at
first like a terminological quibble, but confusing our methods with our
methodologies with our epistemologies can, and often does, put unreasonable
limits on our ability to think creatively and productively about using methods
and about building the methodologies that will most effectively address our
research questions.

. Finally, what of theory, the thing that provides actual
explanatory/interpretive/political content to our attempts to address our
questions about international organizations?63 Theory is in fact orthogonal not
only to method but to methodology and epistemology as well. Take as an example
two theoretical constructs frequently used to study international organizations,
neoliberal institutionalism (understood as the application of Coasian logic to
IOs) and gender analysis (understood as the study of the effects of gender as a
social or discursive construct on the practice of international organization). A
neoliberal lens is compatible with, among other things, formal modeling (to
determine what should happen in the abstract), statistical modeling (to
determine what tends to happen), and discourse analysis (to determine how the
discourse of neoliberalism affected regime shape in particular historical
settings). A gender lens is similarly compatible with various kinds of discourse
analyses, but is also compatible with various kinds of statistical analyses, for
example in answer to questions about what kinds of roles women play in IOs.

. Professionals, understood broadly as individuals who earn a living from their
activities, are central protagonists in all forms of IO action, from diplomatic
encounters to the design, implementation, and evaluation of international public
policies. Though UN diplomatic and elite bureaucratic staff have received some
attention in recent years,7 IO professionals remain relatively understudied as a
population. Capturing diversity and hierarchy within IO professional ranks, a
survey can provide a macro-level description of an organization’s workforce—the
scope of which depends on sampling. Though not exhaustive, possible variables
include: (1) personal information (i.e., nationality, age, civil status); (2)
socioeconomic status (i.e., income, education, occupation); (3) professional
experience (i.e., previous jobs, contractual status, job satisfaction); and (4)
ideational values and preferences (i.e., economic/political/religious/moral
attitudes). Additionally, datasets can be compared and contrasted to assess the
(non)specificity of IO professionals in relation to other populations. For
example, to understand the social Page 71 →construction of international
peacebuilding, Goetze (2017) used web-based survey (WBS) methods to describe the
experiences, professional trajectories, political values, and working
environments of UN peacebuilders.

. Choosing which questions to include in a survey depends on the core research
focus. Based on our research objectives, our questionnaire incorporates two main
types of questions: 1) questions specific to the humanitarian field—i.e., What
is your main motivation for working in the humanitarian sector? How many years
have you worked in the humanitarian sector?; and 2) general questions about
participants’ socioeconomic status, beliefs, and values—i.e., How important is
it for you to live in a country that is governed democratically?. Regarding the
latter, we have drawn questions from the World Values Survey (WVS) and the
European Values Study (EVS), both recognized for their scientific rigor—a
strategy also deployed by Goetze (2017). Within both surveys, we sampled
questions with a demonstrated predictive and convergent validity over time,
meaning they were replicated several times from 1981 to 2008 to control for
survey-related biases such as social desirability8 or recall bias.9

. Choosing which questions to include in a survey depends on the core research
focus. Based on our research objectives, our questionnaire incorporates two main
types of questions: 1) questions specific to the humanitarian field—i.e., What
is your main motivation for working in the humanitarian sector? How many years
have you worked in the humanitarian sector?; and 2) general questions about
participants’ socioeconomic status, beliefs, and values—i.e., How important is
it for you to live in a country that is governed democratically?. Regarding the
latter, we have drawn questions from the World Values Survey (WVS) and the
European Values Study (EVS), both recognized for their scientific rigor—a
strategy also deployed by Goetze (2017). Within both surveys, we sampled
questions with a demonstrated predictive and convergent validity over time,
meaning they were replicated several times from 1981 to 2008 to control for
survey-related biases such as social desirability8 or recall bias.9


PREVIOUS CHAPTER

Footnotes


NEXT CHAPTER

Footnotes

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0
International License.


SHARE

FacebookTwitterEmailPrintשתף



FOOTNOTES


ADDITIONAL INFORMATION

ISBN
9780472903542
Related ISBN(s)
9780472056224, 9780472076222
MARC Record
Download
OCLC
1378527995
Launched on MUSE
2024-01-28
Language
English
Open Access
Yes
Creative Commons
CC-BY-NC


COPYRIGHT

2023




PROJECT MUSE MISSION

Project MUSE promotes the creation and dissemination of essential humanities and
social science resources through collaboration with libraries, publishers, and
scholars worldwide. Forged from a partnership between a university press and a
library, Project MUSE is a trusted part of the academic and scholarly community
it serves.


About
 * MUSE Story
 * Publishers
 * Discovery Partners
 * Journal Subscribers
 * Book Customers
 * Conferences

What's on Muse
 * Open Access
 * Journals
 * Books
 * The Complete Prose of T. S. Eliot
 * MUSE in Focus


Resources
 * News & Announcements
 * Email Sign-Up
 * Promotional Materials
 * Presentations
 * Get Alerts

Information For
 * Publishers
 * Librarians
 * Individuals
 * Instructors


Contact
 * Contact Us
 * Help

 * 1. 
   2. 
   3. 

Policy & Terms
 * Accessibility
 * Privacy Policy
 * Terms of Use



2715 North Charles Street
Baltimore, Maryland, USA 21218

+1 (410) 516-6989
muse@jh.edu

©2024 Project MUSE. Produced by Johns Hopkins University Press in collaboration
with The Sheridan Libraries.

Now and Always,
The Trusted Content Your Research Requires



Now and Always, The Trusted Content Your Research Requires

Built on the Johns Hopkins University Campus




Built on the Johns Hopkins University Campus

©2024 Project MUSE. Produced by Johns Hopkins University Press in collaboration
with The Sheridan Libraries.

Back To Top
✓
תודה על השיתוף
AddToAny
More…