Location, location, location: Measuring public opinion in Ukraine’s regions

The following is a guest post from Ralph S. Clem, emeritus professor of geography at Florida International University, Miami.
*****
As the crisis in Ukraine continues to unfold, and the presidential election scheduled for May 25 approaches, it is essential that both policymakers and the interested public have as much information as possible about the attitudes of that country’s citizens concerning the central issues involved: the legitimacy of the interim government, the Western or Eastern orientation of foreign and economic ties, the extent to which Russia should be involved in Ukraine’s affairs, and, most importantly, whether or not parts of the country should be allowed to secede. The best source of that kind of information is public opinion polling, at least those polls that follow accepted practices with regard to coverage and sampling. That being said, it is less obvious but also crucial that the collection and subsequent release of polling data be done in a geographical context that takes into account the spatial unevenness in attitudes that is almost always evident in matters political (more so in large and internally diverse countries such as Ukraine). Put another way, aggregating data into relatively large geographical units obscures the very place-specific information that we need to appreciate the reality on the ground and the implications thereof, including just how significant the upcoming election will be.
An excellent example of how an otherwise well-designed and scientifically sound poll can produce a potentially misleading impression of political and geopolitical views in Ukraine is the recently released and very widely cited Pew Spring 2014 Global Attitudes Survey for Ukraine and Russia. As will be shown here, this is because of the manner in which the results of the Pew survey are portrayed geographically. To illustrate the point, I refer specifically to the question in that survey that has garnered the most attention and headlines, the one that asked respondents in Ukraine whether or not Ukraine should remain united, which is, arguably, the crux of the matter at hand. But the problem of using overly gross geographical units applies to the other questions and responses discussed in that report as well.
In a somewhat reassuring tone, the Pew document (titled “Despite Concerns about Governance, Ukrainians Want to Remain One Country”) states that: “A clear majority of Ukrainians agree that their country should remain a single, unified state …” The report goes on to describe results from the question on a unified Ukraine (and other questions) in a spatial format of Ukraine divided into three regions: West, East and Crimea, with those regions defined by this map.
Pew Survey Regions of Ukraine (Pew Research Center)
By now even the casual observer of the ongoing troubles in Ukraine will have some notion of an “East-West divide” in that country with regard to politics and socioeconomic characteristics of the population because it has been so extensively depicted that way. Further, most who follow events in Ukraine will know that the present difficulties are most problematic in the “East” (although how “East” is defined varies from one article or report to the next). So it is not unreasonable to use this simple bifurcation (with the de facto seceded Crimea considered separately) as a way of describing very generally how Ukrainians view issues related to the present crisis. Here is the Pew national and regional breakdown for the united (or not) Ukraine question:

With an appreciable majority of 77 percent in favor of unity countrywide, the report notes that “A majority of east Ukraine also wants to be one country (70 percent)”, which seems to be the most noteworthy finding of this particular survey.
With the report’s title in mind, ask yourself this question: If the clear majority of Ukrainians want their country to remain intact, and even the problematic East appears to go along with that idea, why are we on the brink of civil war in Ukraine, or worse, at this writing? The answer is that it is very likely not the case that in some parts of the country — especially certain areas in the “East” as used by Pew — anywhere approaching 70 percent of the population prefers a united Ukraine extant, and that critical fact is not captured by the Pew results as they have been presented.
The fundamental problem is that the overall 70 percent of those polled in Pew’s “East” Ukraine who say that they want Ukraine to remain united is not a meaningful number because Pew’s “East” Ukraine is so large and varied that it is not meaningful as a unit of analysis in any practical sense having to do with the determinants of political attitudes, especially now. The “East” region that Pew uses here comprises 11 of Ukraine’s 24 provinces  (or “oblasts,” not including Crimea and the former federal city of Sevastopol; the city of Kyiv has a separate federal status not included with Kyiv Oblast). Thus defined, the “East” accounts for just over half of Ukraine’s total population and area; by comparison, this gives the “East” a population larger than all of Scandinavia and an area larger than Italy. Put another way, if Pew’s “East” Ukraine were an independent country, it would be the ninth largest and the 10th most populous in Europe.
Not surprisingly for a region this large, there are significant internal variations within the “East” itself with regard to such key socioeconomic measures as the level of urbanization and per capita income, both of which correlate with political opinion, not to mention differences in ethnic identification and language affinities. In fact, just about any socioeconomic or political variable that one can imagine rises or falls along an axis running northwest to southeast from roughly, say, Lviv, to Luhansk. Further, if data are disaggregated into smaller spatial units, it becomes clear that change along this axis follows a gradual gradient, belying the idea that there is some kind of clearly defined or dramatic divide between “West” and “East” Ukraine. A good example of how this looks is the map below compiled by the Ukrainian electoral geographer Serhiy Vasylchenko showing the percentage of the vote won by the now deposed Ukrainian President Viktor Yanukovych in the pivotal 2010 presidential election. Note especially how much difference there is even within the “East,” with vote shares for Yanukovych ranging from the low 20s in Sumy Oblast to the 80s and even 90s in Donetsk and Luhansk oblasts at the eastern end of the axis. The vote for Yanukovych’s opponent, Yulia Tymoshenko, would be the mirror opposite, i.e., she had her highest vote shares in the West and dropped to single digits at the eastern border.

Apropos of this last point, a survey conducted by the Kyiv International Institute of Sociology (KIIS) focusing on “South-East Ukraine” done about the same time as the Pew survey gives us some similar questions and answers at the oblast level for eight of the 11 units within Pew’s “East” and arrives at almost an identical regional figure (69.7 percent) as did Pew for the percentage of respondents who said either absolutely not or probably not in response to a question about whether one’s oblast should secede from Ukraine and join Russia. The KIIS study, however, because it uses a finer geographic scale to depict responses, shows how much intra-regional difference there is between places like Kherson and Luhansk oblasts (both within Pew’s “East”). For example, about 85 percent of respondents in Kherson Oblast said absolutely or probably not to the secession question whereas in Luhansk just barely a half (52 percent) said the same. If we had the Pew data disaggregated even to the oblast level we would probably see the same thing.
This is especially frustrating because, frankly, what we really care about right now are Donetsk and Luhansk oblasts, the locus of the armed revolt against the interim government and the areas where even holding the May 25 vote will be most challenging, as detailed in a recent Monkey Cage blog by Erik Herron and colleagues (note again the northwest-southeast trend line in their maps using small areal units). Secondly, it is also important to know that other key areas in Pew’s “East,” such as Dnipropetrovsk and Kharkiv, are almost certainly relatively more likely to want to remain in Ukraine (as is evident in the KIIS data), meaning that the secession “contagion” might be limited to Donetsk and Luhansk, but we cannot tell this either because it is buried in Pew’s “East” aggregate figure. Also, as has been the case previously, some of these Eastern oblasts may turn out to be pivotal swing regions in the May 25  election.
Whether or not the problems evident in Donetsk and Luhansk derive from underlying socioeconomic characteristics, elite and oligarchic interests, or just the fact that they physically adjoin Russia and make it easier for cross-border infiltration of arms and provocateurs (noting that another Eastern Ukraine border oblast, Sumy, is quiet), or some combination of these, our understanding of the internal political dynamics of Ukraine would be enhanced by survey data disaggregated to the smallest geographical units possible given the limitations of sample size to complement the results of the crucial May 25 election.
*****
Past Monkey Cage posts on developments in Ukraine, Russia and Crimea can be found by clicking here.  Recent posts include:
Erik Herron: Is Ukraine ready to vote?
Joshua Tucker: Despite problems, Ukrainians favor unity
Kimberly Marten: Ukraine and the problem of local warlords
Dmitry Gorenburg: Opposition in Ukraine: A tale of two maps
Kyle Dropp, Joshua D. Kertzer and Thomas Zeitzoff: The less Americans know about Ukraine’s location, the more they want U.S. to intervene
[wapoad type=”inline”]