Introduction to the Special Topic Forum: Using Archival and Secondary Data Sources in Supply Chain Management Research
Calantone, Roger J., Vickery, Shawnee K., Journal of Supply Chain Management
Much empirical research in supply chain management has been based on the analysis of primary data--data that have been collected for the purposes of the research at hand. Typical methods for collecting primary data include research surveys/questionnaires, direct observations and case interviews. In contrast to primary data, secondary data are data that exist for some other purpose than the research at hand. Generally, any database in existence before the initiation of a research study is data that are secondary to that study. The reasoning is that conceptualization of the original data collection design was entirely independent of any consideration of the current study. The most obvious source of secondary research data for business researchers is the public reporting of firms, which is generally undertaken to satisfy a variety of regulations emanating from various government sources. A researcher might obtain secondary data from surveys conducted by other researchers or research organizations, from survey results published in the literature and from existing archives developed and maintained by companies, universities, industry organizations, private/public agencies and/or government agencies.
Archival materials are often unique, irreplaceable, one-of-a-kind items that cannot be obtained elsewhere (Hill 1993, p. 22). Nevertheless, archival data are ubiquitous in our society and world. For example, the U.S. government collects and maintains data related to the census, labor statistics and housing starts. Public and private libraries are significant repositories of archival data including historical records relating to individuals (letters, papers, diaries, computer files, financial records) as well as to corporations. Industry organizations often collect and track data related to key industry parameters and performance. Universities maintain archives related to pertinent areas of expertise including databases such as University of Michigan's well-known Inter-University Consortium for Political and Social Research (ICPSR) (Kiecolt and Nathan 1985, p. 19). Company archives are also important sources of data in the form of business records, memos, administrative files, e-mails and official correspondence. Generally, archival data are the original and historical record of activity of many sorts, preserved for purposes that may or may not have research implications immediately (e.g., sentimental value).
Primary data have a reliability advantage in the sense that the researcher knows where it came from and how it was collected. Nevertheless, there are several advantages to using secondary data in empirical studies. First, secondary data are more publicly available to a large number of scholars allowing for true re-search, as well as replication and validation studies. Second, secondary data sourced from archives can be more objective than even primary survey data because it is free from contamination by respondent perceptions and/or memories of the phenomenon of interest. Indeed, some secondary data (such as the financial performance figures and ratios produced by Compustat) can be classified as purely objective data. A third advantage is that researchers can use secondary data from surveys, censuses, etc., to answer questions or test hypotheses that are far removed from the research intentions or informational requirements of the scholars and/or agencies whose studies or initiatives generated the data. This enhances the credibility of the new research because it removes the possibility that the purpose or intention of the research could have influenced the design of the research questions, survey instrument and/or population(s) sampled. Summarily, the intrepid researcher may avoid biases from both source and researcher by the use of secondary data alone or in combination with primary data.
Secondary data research requires less money, less time and fewer personnel. Furthermore, a variety of specialized tasks can be completed before a new survey collection, such as defining which groups need over sampling, which research questions need elaboration, which hypotheses should be revised and whether there is a need to refine measures (Kiecolt and Nathan 1985, p. …