There is also an approach of researching a variety of sub-national cases with an aim at national level generalisation, which is most useful for cross-national comparison that seeks to get beyond crude country-level scoring. Broad surveys and single case studies are better suited to the investigation of questions at later points in the research cycle, when hypothesis-testing and specification of fine-grained causal mechanisms are of primary importance, while sub-national comparison is best for questions about which hypotheses are still being formulated.
Sub-national comparison can be useful for specifying hypotheses and the antecedent conditions required for them to operate, thus telling us how general specific findings may be and where they can be most usefully tested. It can also help us build what has been called ‘typological theory’, by categorising sub-national level cases along one or more dimensions. Finally, it can help us evaluate some of the possible causal mechanisms lying behind observed relationships.
For example, we can reach some tentative understandings of where state enterprise lay-offs took place and for which reasons in the late 1990s and early 2000s using sub-national comparison. While workers lost jobs in the Northeast due to long-standing economic difficulties, in the Central Coast unemployment was mandated by government fiat. Thus, we should look in further research for evidence of these processes in their respective regions.
Such evidence could result from quantitative analysis of survey data or other indicators to test the basic causal relationships. Alternatively, the evidence could come as a product of in depth site intensive research or case study research to tease out more fine-grained details of the causal mechanisms in play.
So how can sub-national comparative analysis be usefully informed and better situated within the terrain of Chinese politics?
When undertaking sub-national comparative analysis, I recommend following four basic steps to ensure maximum intelligibility and applicability beyond the specific research question, while retaining accuracy. First, researchers must pay close attention to how they define sub-national populations and select cases for analysis. Second, researchers need to formulate clear ideas very early in the conduct of their research of what representative cases within each population might look like. Third, care should be taken to avoid over-extension and under-estimation of theories. Finally, researchers must remain focused on illuminating important aspects of Chinese politics as a whole.
It has been usual to define sub-national populations and cases as coterminous with sub-national political units (like Chinese provinces). This method is only suitable for some research questions. By defining populations in ways tailored to the substance of research and justifying this process up-front, the researcher is both more transparent about methodology and more likely to avoid a number of pitfalls.
By knowing and explaining clearly what makes their cases representative, authors make their research both more robust and much easier for readers to follow and digest. While it may be impractical to set out for fieldwork in China with a predetermined set of cases, knowing which ones might be representative of which populations is crucial. Explaining this to the reader is also necessary to place one’s findings more clearly in context.
By specifying the range of cases to which their ideas can be properly applied, authors make it easier for readers to evaluate their claims fairly and appropriately. They also avoid over-generalising all of China as one atypical city or village and at the same time steer clear of underestimating the breadth of their insights.
Keeping China as a whole in their sights enables scholars to speak to issues of intrinsic interest to China specialists and those most useful to generalist social scientists. By saying something informative about China at the national level, researchers address the broadest possible audience of China experts and make their work most useful to cross-national comparativists.
William Hurst is Assistant Professor at University of Texas at Austin. He specialises Chinese labour politics and on the political analysis of courts and legal institutions in China and Indonesia.