Keman, Hans. "Comparative research methods". In Comparative Politics, edited by Daniele Caramani, 47-60, Oxford: Oxford University Press, 2014.

Keman, Hans. "Comparative research methods". In Comparative Politics, edited by Daniele Caramani, 47-60, Oxford: Oxford University Press, 2014.

• Comparative cases exist in many different forms. The five most common forms are multiple case studies, including multiple cases with the same variables; time series, of one case over time; closed universes, some cases in different relevant periods; cross sections, of many cases during one time period; and pooled analysis, looking at cases regardless of time (52-53). This categorization was initially developed by Drs. P. Pennings, H. Keman, and J. Kleinnijenhuis in the second edition of their 2006 book on research in political science.
• The basic method of comparative politics is taken from John Stuart Mill's 1843 publication, A System of Logic, in which he notes that studies of politics either look towards explaining differences or similarities. So either studying differences in otherwise similar cases, or studying similarities in otherwise disparate cases (54-55).
• A consistent issue in the design of experiments and studies in political science is setting the limits for categories. For example, a study of a federal state first requires defining what constitutes federalism, a difficult task open to interpretation because empirical realities rarely match mathematical categorization (56). This point was initially raised by Dr. J. Van Deth in his 1998 book on comparative politics.
• Dr. Giovanni Sartori developed a graph to demonstrate the risks inherent in political science categories. The larger a data set is and the more a concept is stretched, the more applicable its findings are, but the less likely it is to be accurate. On the converse, the more limited a set, the less applicable its findings, despite their accuracy (56-57).
• A number of fallacies are present in the interpretation of comparative political science. The most common are Galton's Problem, a failure to account for exogenous factors; individual and ecological fallacies, where aggregate data is applied to individual choices or visa versa; and selection bias in the countries and variables studied (57-58).