Sampling on the Dependent Variable: French Peasant Edition
25 October 2009 at 5:40 pm Peter G. Klein 4 comments
| Peter Klein |
A useful example of the methodological flaw that plagues the “great companies” and “great leaders” literature in management, from Graham Robb’s excellent The Discovery of France (Norton, 2007):
[N]early every autobiographical account of ordinary life in eighteenth- and nineteenth-century France comes from the early chapters of memoirs written by exceptional men who rose through the ranks of the army or the Church, woo wrote their way to fame or who were plucked from obscurity by a patron, a lover or, eventually, an electorate. Few men and even fewer women had the means or the desire to write a book on “How I failed to overcome my humble origins.” Apart from the countless riches-to-riches tales written by aristocrats, almost all the lives that we know about follow the same untypical upward trend: the farmer’s son Restif de la Bretonne, the cutler’s son Diderot, the watchmaker’s son Rousseau, the Corsican cadet Napoleone Buonaparte.
These spectacular success are more typical of long-term trends than of individual lives. Categorical terms like “peasants,” “artisans,” and “the poor” reduce the majority of the population to smudges in a crowd scene that no degree of magnification could resolve into a group of faces. They suggest a large and luckless contingent that filled in the background of important events and participated in the nation’s historical development by suffering and engaging in a semblance of economic activity.
Likewise, business and entrepreneurial strategies can be understood by studying not only firms that tried them and succeeded, but also those that used the same strategies and failed. Reducing the majority of companies to smudges in an industry-wide or economy-wide crowd scene tells us little about what does and doesn’t work.
Entry filed under: - Klein -, Management Theory, Methods/Methodology/Theory of Science.
4 Comments Add your own
Leave a comment
Trackback this post | Subscribe to the comments via RSS Feed









1.
Steve Phelan | 26 October 2009 at 12:01 am
Excellent point, Peter!
2.
Bo | 26 October 2009 at 11:55 am
Need we make a reference here to the recent visit of Oprah to Denmark and her excellent 22 minutes account of how the most happy people on earth live….
3.
RD | 13 February 2013 at 5:57 am
Sampling on the dependent variable is not necessarily flawed, if you use a case-control design. And it can be extremely useful, especially when studying rare events. See: http://soq.sagepub.com/content/10/3/269
4.
Mjavad | 13 April 2013 at 12:52 pm
An excellent piece by lovely Clay Christensen elaborated on the point almost ten years ago:
http://hbr.org/2003/09/why-hard-nosed-executives-should-care-about-management-theory/ar/1