3
they would be available for the entire 1989-1999 period. Comparing the results
of disparate surveys poses additional methodological problems. Unfortunately,
panel studies on ethnic politics have not been carried out in either Latvia or
Ukraine.
1
The easiest source of quantitative data on nationalist mobilization is
electoral data showing the vote for nationalist parties in various districts across
time. The recourse to aggregate data, however, risks committing the so-called
“ecological fallacy” (Robinson 1950; Achen and Shively 1995; King 1997).
Multiple individual-level behaviors may be consistent with the same macro-level
outcome. The ecological fallacy does not necessarily pose a significant threat in
the case of nationalist parties. For example, we can assume that differences in
the Latvian nationalist vote across districts reflect the varying support for
nationalism among ethnic Latvian voters. We do not suppose that large numbers
of non-Latvian voters cast ballots for Latvian nationalist politicians (Bloom 1999).
Several specific problems surface with the use of electoral data to
measure nationalist mobilization and demobilization in Latvia and Ukraine. First,
election results give at most snap-shots of the preferences of voters, due to the
infrequency of elections. Second, electoral data from Ukraine is problematic in
the pre-1998 period because non-partisan candidates dominated in single-
1
Several surveys of interest have been conducted in Latvia and Ukraine. David Laitin and Jerry
Hough carried out surveys between 1993 and 1995 in Kazakhstan, Estonia, Latvia and Ukraine.
See the results in Laitin (1998: 201-16). The Latvian Naturalization Bureau commissioned
interesting surveys of the Latvian citizen and non-citizen populations in November 1997 and
January 1998 (Baltijas datu nams 1998). The Institute of Sociology of the Ukrainian Academy of
Sciences produced a series of representative surveys of the Ukrainian population in 1994, 1995,
1996 and 1997. See Holovakha and Panina (1998 ) for the results of these surveys.
Convention