Lost in Aggregation: Improving Event Analysis with Report-Level Data Academic Article uri icon

abstract

  • AbstractMost measures of social conflict processes are derived from primary and secondary source reports. In many cases, reports are used to create eventlevel data sets by aggregating information from multiple, and often conflicting, reports to single event observations. We argue that this preaggregation is less innocuous than it seems, costing applied researchers opportunities for improved inference. First, researchers cannot evaluate the consequences of different methods of report aggregation. Second, aggregation discards reportlevel information (i.e., variation across reports) that is useful in addressing measurement error inherent in event data. Therefore, we advocate that data should be supplied and analyzed at the report level. We demonstrate the consequences of using aggregated event data as a predictor or outcome variable, and how analysis can be improved using reportlevel information directly. These gains are demonstrated with simulateddata experiments and in the analysis of realworld data, using the newly available Mass Mobilization in Autocracies Database (MMAD).

published proceedings

  • AMERICAN JOURNAL OF POLITICAL SCIENCE

altmetric score

  • 10.4

author list (cited authors)

  • Cook, S. J., & Weidmann, N. B.

citation count

  • 12

complete list of authors

  • Cook, Scott J||Weidmann, Nils B

publication date

  • January 2019

publisher