The use of secondary data in purchasing and supply management (P/SM) research

https://doi.org/10.1016/j.pursup.2016.08.005Get rights and content

Highlights

  • Secondary data is growing in importance in purchasing and supply management research.

  • Secondary data can be a principal of supplemental data source.

  • Different sources of secondary data have different advantages and disadvantages.

  • Big data is an increasingly important, generally unstructured form of secondary data.

Abstract

The use of secondary data, data that has been gathered for another purpose but may be suitable for research, is growing in relevance and importance in purchasing and supply management research. This paper presents issues associated with using secondary data as a principal source of data or as supplemental to other data or methods. As editors increasingly call for multiple data sources in research publications, secondary data become even more relevant. Secondary data have many benefits, and some secondary data sets are well established and highly credible. The authors describe some of the ways that secondary data has been used in purchasing and supply management research, as well as some of the sources of data. Secondary data does have its limitations, ranging from bias in collection and reporting to difficulty in identifying and accessing appropriate sources of secondary data to dealing with its predominantly unstructured nature. Suggestions for improving the reliability and validity of secondary data are provided, as well advice for dealing with big data.

Introduction

Secondary data is defined as quantitative or qualitative data that has been collected by someone other than the researcher(s) for a different purpose than its intended use in research. There are many different types of secondary data available. Some of the more commonly used are existing literature, census data, governmental information, financial data, organizational reports and records (Lind et al., 2012). This data may be free to access, available by permission of the party collecting the data, or may require payment of a fee.

Researchers in P/SM often use secondary data to triangulate findings from principal data collection such as interviews, case studies, surveys and experiments (Tatsis et al., 2006, Sancha et al., 2015). One common use is as an objective source of performance outcomes associated with various P/SM practices among a research sample. Researchers have shown that the use of secondary data as the principal data source can reduce the bias that is sometimes introduced during case studies and the intrusiveness of data collection that is inherent in more experiential methods such as action research, experiments or interviews (Rabinovich and Cheon, 2011). Secondary data can also help overcome issues of survey fatigue or the use of survey companies such as QDAMiner or survey monkey that have questionable demographics and a lack of control over survey respondents (Schoenherr et al., 2015).

There are some important practices for using secondary data in order to address the appropriate research questions, which are discussed in this note. With the increased awareness of data availability, calls for the use of archival and secondary data as a principal source of data in P/SM research are increasing, as there are numerous benefits to its use (Calantone and Vickery, 2009).

Section snippets

Why use secondary data in P/SM research

Increasingly, it has been difficult to garner significant response rates using a survey method, and the litigious propensity of the current society has made access to case study participants challenging. Many populations have been over-sampled, and suffer from survey fatigue, which creates an overall unwillingness to respond to surveys. In order to gather survey data, researchers often rely on the expensive services of survey companies such as survey monkey and Qualtrics that have developed

Benefits of using secondary data

There are many potential benefits to using secondary data, as shown in Table 1 below. For example, there are many sources of such data. Some are available free through libraries and other sources, and others may be purchased at a relatively low price versus what it would cost for a researcher to create the dataset. This may save a great deal of time and human effort, although some formatting and cleaning of secondary data sets is often required.

Secondary datasets often use well-established

What type of data has been used in P/SM research?

For illustrative purposes, a search on google scholar with the keywords “secondary data” and “supply” was run for the years 2010 until 2015. The first 200 hits were checked for potential applicability based on the brief abstract, the key words and the journal where the article was published. Of these 200, there were 62 articles that warranted full paper reviews. Of these 62 articles, 21 used secondary data as principal data sources and 20 used secondary data to triangulate findings from other

Why does secondary data make sense for P/SM research?

There are many reasons why secondary data is a good fit for P/SM research. If using a broad, well-established source of data, there are fewer chances to skew the data collection process based on researchers preconception and bias. This is essential to generating meaningful, generalizable and publishable results.

Further, secondary data sets are often already validated. This allows researchers to focus on validation of new constructs and measures that are critical to move research forward.

General limitations of secondary data

When seeking secondary data to use in analysis, there can be significant time spent searching for appropriate datasets and interpreting the data, trying to understand issues such as: where are appropriate data published, are there common reporting measures used for this type of data and are there common time frames? Also, a significant problem is that the a single database or source secondary data may not be available to answer the research question, so researchers must be able to appropriately

Specific limitations of voluntarily reported data

Specific to voluntarily reported data such as CSR data and non-financial data from company websites, there is no auditing of the data. While it is unlikely that companies would wildly distort such data due to the potential impact that this could have on reputation, self-reported data tend to be presented in a favorable light (Jose and Lee, 2007). In addition, because there is no requirement that companies disclose everything, reports will likely emphasize what is perceived as important, and

What are some techniques to ensure reliability and validity?

Reliability assesses whether the data is consistently reported over time. With secondary data, it is thus essential to understand how the data have been collected. Were consistent measures used over time, and across different data sets purporting to measure the same thing? The researcher should be able to explain what was being measured, how, and in some cases by whom (self-reporting versus layperson reporting, versus professional researcher). This helps ensure that differences in the data are

Where does big data fit in?

Big data is a term used to describe the massive amounts of data that organizations are collecting that cannot be processed, analyzed and managed, using traditional tools and approaches (Zikopoulos et al., 2012). Big data is formally defined as data that has volume, velocity, variety, and veracity (IBM, 2016). To the organization collecting it, it is primary data. When we access it as researchers who did not collect the data or direct the collection of the data for our purposes, it becomes

Future research opportunities using secondary data

With the growth of “big” data, the number of opportunities for using secondary data is increasing. The important consideration is ensuring that the data is used in a way that is meaningful and adds value in addressing real problems and issues, not simply in a way that is convenient. Using traditional secondary data sources such as the COMPUSTAT databases, census databases and others will continue to be important. More companies have massive quantities of data that they capture on their

Conclusions

While there is a wealth of potentially valuable and interesting secondary data that can be mined by researchers, including a growing store of mainly unstructured big data, the acquisition, cleansing, interpretation and publication of that data is not without extreme challenges. Reviewers may not be familiar with such data. This creates an extra responsibility for the authors to take care in the handling of the data, its interpretation and its presentation to the public. As in all data analysis,

References (34)

  • N. Altay et al.

    Impact of disasters on firms in different sectors: implications for supply chains

    J. Supply Chain Manag.

    (2010)
  • C. Busse

    A procedure for secondary data analysis: innovation by logistics service providers

    J. Supply Chain Manag.

    (2010)
  • R. Calantone et al.

    Special topic forum on using archival and secondary data sources in supply chain management research

    J. Supply Chain Manag.

    (2009)
  • M. Crum et al.

    Sustainable supply chain management: evolution and future directions

    Int. J. Phys. Distrib. Logist. Manag.

    (2011)
  • A. Ellinger et al.

    The influence of supply chain management competency on customer satisfaction and shareholder value

    Supply Chain Manag.: Int. J.

    (2012)
  • A.E. Ellinger et al.

    Supply chain management competency and firm financial success

    J. Bus. Logist.

    (2011)
  • L.M. Ellram et al.

    Factor-market rivalry and competition for supply chain resources

    J. Supply Chain Manag.

    (2013)
  • Cited by (0)

    The purpose of this note is to discuss the use of secondary data in purchasing/supply management (P/SM) research. Secondary data is widely available and may be useful as a principal data source or as supplemental data for other research methodologies.

    View full text