Introduction: Research and Methodologyeliminated because the alliance negotiator, Bismarck, intention-
ally failed to sign the final treaty.^3
Second, several alliances were identified as meeting the
Singer and Small definition, but the actual treaty text could not
be located. For example, the terms of a few recent alliances were
confirmed by multiple press accounts, but the actual treaty texts
have not been deposited with the United Nations or otherwise
released in full to the public. The texts for a few nineteenth-cen-
tury treaties have not been reprinted, were never cataloged in
the various treaty collections, and should probably be consid-
ered lost. Alliances with no surviving treaty texts were included
in this data set only when their existence and characteristics
could be confirmed by at least two secondary sources. This deci-
sion rule adds only one alliance before 1945—an alliance
between France and Sardinia in 1859. In all, forty-one other
alliances are included on the basis of secondary sources only; all
were signed after 1945, with more than half formed after 1980.
Sources Used
The generation of this data set began with the various data
sources currently available—the original Correlates of War data
set and the various independent updates. These data sources
were augmented with new United Nations data and several new
anthologies that were released after the last revision of the Cor-
relates of War data. In addition, Lexis-Nexis, Keesing’s, and
Facts on File were used for every year available (the years vary
according to each source) to update the twentieth- and twenty-
first-century alliances.
Secondary historical sources were used to collect back-
ground information on the duration of each alliance (exact
abrogation or termination dates), and, as mentioned above, sec-
ondary sources were also used to identify the existence of a for-
mal alliance when no treaty texts could be identified. For more
than 95 percent of the alliances identified using secondary
sources, at least one secondary source contained excerpts of the
alliance; thus, confirmation of the terms was consistent with
treaties for which the full text was available.^4 Once again, in
emulation of the Singer and Small data collection procedures,
secondary sources for these volumes were chosen carefully to
ensure global coverage during the entire time period.^5
Alliance Sources and Avoiding Missed Alliances
Researchers at the Correlates of War project have always
believed that using multiple sources from a broad array of geo-
graphic and historical perspectives is the best way to try to iden-
tify the universe of formal alliances (Singer and Small 1966).
Unfortunately, it is not easy to avoid omissions in this type of
research because alliances may exist that were never registered,
were never included in anthologies, and were never mentioned
by historians. The prospect of these missing alliances is trou-
bling, but the vigorous data collection procedures used give an
accurate representation of the universe of alliances during the
years since 1816.
As a reliability test of the methods used to collect these
alliances, twenty state-year periods of ten years were chosen
randomly and were then independently investigated by several
coders not involved in the original construction of the data set.
These in-depth investigations also focused on secondary
sources not originally used in the construction of the data set.
In all but one of these twenty investigations, the alliance portfo-
lio produced by these more intense investigations matched the
original alliance portfolio. In the only aberrant investigation, a
single alliance was added to the current data set. Given these
reliability results, one can be fairly confident that the prospects
for Type I errors in this data set are low.
The original Correlates of War data collection was compiled
using the League of Nations Treaty Series,Britain’s Foreign and
State Papers,as well as various historical monographs. To avoid
subjecting the data collection to particular geographic or cul-
tural biases, researchers included samplings of works from sev-
eral different historical schools and from several different areas
of the globe (Singer and Small 1968, 1–2). Maintaining this
extreme effort to reduce bias was not necessary with the data
extension, however, because of the recent compilation of the
Consolidated Treaty Series.This series, edited by the late Clive
Parry, presents a complete collection of printed treaties since
1648 in their original languages and with French and English
translations where they exist. Parry (1978, preface) included
treaties regardless of where they were signed or which states (or
non-states) were the signatories.
Although the Parry collection (1978) does represent the uni-
verse of written treaties from 1648 to 1851, the accuracy of each
alliance found in the Parry collection has been verified using
external sources. The original sources, cited by Parry, were con-
sulted to verify the authenticity of the treaty text, and indepen-
dent historical monographs and chronologies—especially
Albrecht-Carrié (1958), Babuscio and Dunn (1984), Dupuy and
Dupuy (1977), and Langer (1972)—were searched for informa-
tion about each document. To qualify as an alliance in the 1648
to 1815 data set, the document had to meet the requirements of
system membership, alliance commitment, and be mentioned
in at least twosecondary sources independent of Parry’s collec-
tion (once again, for most of the alliances these sources were
Langer [1972], Albrecht-Carrié [1958], and Dupuy and Dupuy
[1977]).Criteria for Identifying Formal Interstate Alliances
As mentioned above, the Correlates of War definition of a for-
mal international alliance has three explicit criteria: first, the
alliance must be signed by two qualified system members; sec-lv- See Pribram et al. (1920) for a description and terms of the proposed
May 6, 1873, treaty with Russia. - In the few cases where excerpts were unavailable, the summary of the
alliance in the secondary sources had to differ in some way. This rule
prevents the perpetuation of erroneous reporting through multiple
sources. - For a list of all sources, see Appendix II of Gibler and Sarkees (2004).