Author(s)
Source
Working Paper, SSRN, 2007
Summary
This paper examines a study and report by Oberholzer-Gee and Strumpf about recording industry sales and file-sharing.
Policy Relevance
The impact of file-sharing has been a contentious policy issue in recent years and the Oberholzer-Gee/Strumpf paper has been influential. Whether the study meets normal standards for care of honesty is important to the discussion of these issues. This dispute has received lengthy coverage in the Chronicle of HIgher Education and Handelsblatt (a leading German business magazine) and heavily downloaded at SSRN.
Main Points
-
O/S were unwilling to make most of their data available, giving different and conflicting reasons at different points in time. I was able to collect the same data as them for three (of the admittedly less important) tests that they perform in their paper. In each case the results I found were at variance with those claimed by O/S. In each case the qualitative conclusion was the opposite of the one drawn by O/S.
-
O/S make numerous factual claims about sales in the US and international sound recording industry, about the role of big box retailers, about the role of substitute products, and about the causes of financial difficulties for record companies. Virtually every factual statement they make in support of their hypothesis was made without any citation to actual data. When actual data were consulted, their claims were found to be incorrect.
-
When authors refuse to make details of their analysis available to other researchers, when they do not give consistent explanations of why they cannot make such details available, when many of their claimed facts are found to be at variance with actual facts, and when replications of their analysis leads to different results, their study should not be given credence. The O/S study is such a study.
-
Their study does not appear to meet normal standards for care or honesty.