This article provides a history of the treatment of observational errors where conditions cannot be controlled to reduce inaccuracies, more specific, a history of the discussion of errors in social statistics. The main focus is on Oskar Morgenstern’s atypical position in this discussion. In contrast to his contemporary social statisticians, Morgenstern took the natural science approach as the ideal standard for dealing with errors. His position, however, is not atypical when compared with natural science perspectives at that time. His view was attuned with the view of logical empiricism of the 1950s on the difference between natural science and social science: because social science is inexact we need experts to ensure that observations are “scientific.” Moreover, in a “hostile” and “secret” world we need experts to assess the accuracy of the observations.

The text of this article is only available as a PDF.
You do not currently have access to this content.