METHODS
Data and Sample
The data for our study were taken from AMJ
articles published between 1963 and 2007. AMJ’s
first issue was actually published in April 1958 (for
historical reviews, see Adams and Davis [1986];
Kirkman and Law [2005]; Mowday [1997];
Schminke and Mitchell [2003]). However, the first
five volumes of the journal contained few empirical
articles, with many issues instead including essays,
reviews, and discussions of management education
issues. We therefore began our review in 1963, so
that our coding of theory building and theory testing
would be based on a larger set of articles. Our
coding covered all issues of every third volume of
the journal. As each AMJ editor serves a three-year
term, our coding therefore included issues from
every editorial term, beginning with Dalton McFarland
(1961–63) and ending with Sara Rynes (2005–
07). We included the first five issues of 2007 in our
review to be as current as possible. In all, the sample
included 16 volumes, 75 issues, and 770 articles
(AMJ volumes included four issues until 1991,
moved to five in 1992, then began including six
issues in 1993). Of those 770 articles, 667 were
empirical articles that could be coded on theory
building and testing. The remaining 103 articles
were methods pieces, introductions to special research
forums, and conceptual articles written before
the launch of AMR in 1976