With respect to analyzing the supply of quality manuscripts, the Board of Governors, through the Journals Committee, examined 10 specific journals for a three-year period that are not within the AOM portfolio, but publish empirical studies (this set included SMJ, OS, JOM, JAP, P-Psych, JIBS, JOB, OBHDP, JMS, and LQ). They chose the three-year time period 2004-2006 in order to have a minimum of five years' of citation history for published articles (the research was conducted in 2012). They then examined every article in those years that was cited (using ISI counts) with a higher frequency than the average for AMJ articles published in the same time period. The 10 specific journals were selected because they are generally among the top 10 outlets in terms of their five-year citation impact factors.
Having identified the highly cited papers in competing journals, they examined each to see if the article was deemed to be a poor fit for AMJ but a potentially good fit for AMD. Although this was somewhat of a subjective judgment, most of the articles they identified rarely or never used the word "theory" based upon an electronic search of the manuscript, and none of these articles stated theory building or testing theory as their primary purpose.
There were roughly 350 articles published in these competing journals that achieved higher citation counts than the AMJ average for that period. Roughly 95 (25%) of those articles were not building new theory and were not theory-driven. Instead of trying to build new theory or make theoretical extensions, the work reflected in those 95 papers made empirical contributions to the literature by (a) answering phenomenon-driven research questions (e.g., studies that focused on performance in multinational alliances but not theories of multinational alliances, or studies that focused on preventing work-related accidents but not theories of accidents, and so on); (b) delivering replications and extensions of critical past findings (e.g., international replications); (c) providing more refined parameter estimation (e.g., via meta-analyses); (d) informing debates within the existing literature with new data; or (e) conducting construct validation and refinement.
On average, the 95 papers were cited roughly 90 times compared to the average of 45 for AMJ articles published during the same time period, and many of the articles had been rejected at AMJ. Although slightly skewed, the distribution of the citation data was largely symmetrical. This suggests that if the 95 papers that were above the AMJ mean were combined with the next 95 papers below the mean, there would be a sample of roughly 190 papers from which one could build a journal that had a mean number of citations that was similar to that found at AMJ (over the studied three-year period).
There seems to be no shortage of articles that fit with the AMD profile, and many of these types of articles wind up having a great deal of impact as assessed via citation counts. AMD will allow AOM to capture some of the work that is now being published by competing journals. Many of these competing journals are published by commercial vendors with no formal ties to any professional society such as AOM, but that draw on our members as authors and reviewers.