Citation

Dulic T. J. Peace Res. 2004; 41(1): 85-102.

Copyright

(Copyright © 2004, SAGE Publishing)

DOI

10.1177/0022343304040051

PMID

unavailable

Abstract

The article provides a critical analysis of the quantitative method used by Rudolph J. Rummel in order to estimate the democide rate for various political systems in the 20th century. The first part shows that the estimates used by Rummel for Tito's Yugoslavia cannot be relied upon, since they are largely based on hearsay and unscholarly claims frequently made by highly biased authors. The second part shows how the data have influenced the mid-estimates Rummel uses for further statistical analysis. A comparison with demographic research in former Yugoslavia shows that Rummel's mid-estimates for Tito's mass killings are much too high and contravene his data for the population deficit in Yugoslavia. The author also criticizes a key assumption in Rummel's method of reasonable approximation', namely, that overestimations tend to be taken out by underestimations. It is shown why such a proposition is problematic, particularly in this case where there is a wide discrepancy between high and low estimates. Although the article concentrates on communist terror in former Yugoslavia, the results may have wider implications for Rummel's research if he uses similar sources in other case studies. If so, Rummel would need to revise the method and exclude unreliable estimations in order to obtain useful data.