CHAPTER 9

Analyzing Secondary Customer Data

In this chapter, we will describe approaches to the analysis of secondary analysis that you can effectively apply in your own marketing efforts. We will look at different kinds of analyses and how you can put consumer data to work for you.

Archival Data

In the last chapter, we largely focused on the re-analysis of data or research reports originally collected for purposes of research. The original research that gave rise to the data or research report may have had a different purpose than that of the intended secondary analysis, but the process that gave rise to it included a research design that when adequately described provides a basis for evaluating the integrity, relevance, and generalizability of the underlying data and reported results. However, there are many other sources of secondary information that did not originate in a formal research process. Government, corporate and church records, reports and filings, personal correspondence, medical treatment records, minutes of meetings, court records, and genealogical records are all examples of written documents that may be archived and used for secondary analysis. However, archives may also include photographs, video and audio recordings, oral histories, maps, furniture, architectural drawings, and land surveys among others. Indeed, almost any artifact can become the basis for archival research.

While it might seem that archival research is most useful for conducting historical analysis, and a great deal of archival research is of a historical nature, archival data also lend themselves to other types of secondary analyses. For example, a very robust stream of research on the economics and sociology of organizational relationships rests on the secondary analysis of legal contracts.¹ Tax records, which are collected for government record keeping and legal purposes, have been used in secondary analyses to evaluate the effects of different tax policies over time and across different political units on such things as philanthropic giving, entrepreneurial activity, and investment in research and development by corporations. The spectrum of research topics that have been addressed by secondary analysis is very broad and limited only be the imagination and creativity of researchers.

The secondary analysis of archival records and artifacts brings special problems, however. Since the underlying data or artifact was not created by a formal research process there is a need to understand the reasons for its existence as well as why it was saved or survived. The context in which the underlying data were created also is important to understanding its meaning. Authenticating the archival artifact can also be a challenge, especially for such things as personal correspondence, written accounts of events, and even visual media such as photographs.

Analytic Techniques

Secondary analysis can take many forms. Generally, the primary determinants of the analysis are the research question itself and the characteristics of the underlying data. Identifying information relevant to a particular research question is only the first step in secondary analysis.

In an age of instant access to information it is easy to become overwhelmed by information. Finding order and meaning in a plethora of information is often difficult, particularly when there are inconsistencies, omissions, and differences in methods among various sources.

A common problem faced by researchers employing secondary analysis is that of combining the findings and conclusions of several sources of information. The synthesis of information is an important skill and was long criticized for its lack of objectivity.²

Secondary analysis can range from simple descriptive reporting of results, to new analyses of underlying data, to efforts to combine different sources of information to construct an answer to a new research question, to efforts to integrate a body of literature to reconcile contradictions and draw conclusions about the presence or absence (and strength) of specific effects or variables. A number of authors have addressed specific approaches to the analysis of secondary sources.³ The most common approaches to secondary analysis are:

We will explore each one in detail.

Descriptive Analysis

Descriptive analysis involves describing the attributes, findings, and conclusions of past research. There is little effort to integrate or interpret the underlying data beyond reporting what was found and perhaps counting or otherwise summarizing various results. For example, a review of surveys  of schools undertaken by different researchers that provides the numbers of students with particular special education needs and the number and types of programs offered to such students would be an example of a descriptive analysis.

Interpretive Analysis

In contrast to descriptive analysis, interpretive analysis seeks to go beyond the data or particular set of findings to develop a larger meaning of the underlying data. For example, in the analysis of survey of schools above the secondary analyst might go beyond mere description to related program availability to classroom success and draw conclusions about the relative efficacy of program types. Interpretive analysis is especially common in archival research in which there is a need to place a particular artifact or set of artifacts within a social, cultural, or economic context. Interpretive research often seeks to identify a larger meaning of an artifact by identifying its social or cultural significance and origins.

Comparative Analysis

Comparative analysis focuses on the identification of similarities and differences across sources and data collection efforts. Comparisons may involve analyses of differences over time or among social groups or regions. In such cases comparative analysis may be combined with primary research in a replication or restudy of the original research to follow up the original sample or to make comparisons with additional groups, settings, or circumstances. An example of comparative analysis is found in Di Gropello, who used secondary data to analyze the effects of decentralization of school management in Central America and compare centralized management systems with decentralized management systems.⁴ Key conclusions of this analysis, which was sponsored by the World Bank, were that decentralized, school-based management models produce greater community involvement and greater effort on the part of teachers while producing learning outcomes as high as in traditional schools despite being located primarily in the poorest and more isolated areas.

Verification

Verification is similar to comparative analysis but with the more limited objective of substantiating prior results. The prior results provide the point of comparison to which new data are applied. For example, in examining the efficacy of a particular treatment for a psychological disorder, secondary analysis of medical records may demonstrate that efficacy is similar to what had previously been demonstrated in more controlled clinical research.

Re-Analysis

Among the more common types of secondary analysis is re-analysis of the underlying data. Such re-analysis might take the form of applying different analytic tools or the addition new variables obtained from other sources. More often, re-analysis involves asking new questions of the data that are different from the questions that gave rise to the original data collection effort. This involves approaching the data in ways that weren’t originally intended by using the data to investigate a different research question, theme, or topic.

Generally, the more in-depth the material, and the more information that exists about the underlying data and how it was obtained, the greater the likelihood that the data’s utility for addressing new questions can be evaluated. For example, data originally obtained for purposes of understanding access to health care among lower socio-economic families might be used in secondary analysis to examine the impact of health access and health problems on the educational achievement of children in these families. This might involve the use of data with respect to educational performance obtained in the original study or the combination of data from the health care study with other secondary data related to educational performance.

Finally, in the context of long-term streams of carefully designed research it is often possible to use differences in the research designs of different studies to examine the relative influence of variables that contribute to the obtained outcomes of results. Such meta-analyses involve the statistical analysis of differences and similarities in both the design and results of different studies to identify the degree to which different effects and different design parameters (such as sample or type of measure) explain the pattern of results obtained across studies. There is a rich literature on the technical details of conducting meta-analysis.⁵ The chapters in this volume on meta-synthesis and meta-analysis also provide useful introductions and more technical information about performing such informative analyses.

Ultimately, the type of secondary analysis that a researcher carries out must be dictated by the research question. In many applied contexts a very simple descriptive analysis may be sufficient. In other situations, the research question may require integration across studies and sources but not at highly quantified and specific level. Secondary analysis that focuses on the testing and development of theory may require the quantitative precision of meta-analysis.

Takeaways

Secondary analysis and archival research are very common in both academic research in the social sciences and in applied research designed to address practical policy and business question. The prudent user of such information will know what information is relevant for a given situation and will select an appropriate source of information. Once you have acquired the research for secondary research, you will take it through one or more forms of analysis. The one you choose largely depends on your purpose for the data. After analyzing the secondary data, you’ll have a clearer picture of what additional questions you need to ask through your own research in order to practically apply what you have learned.

Notes

1. MacNeil (1978, pp. 854–905).

2. Glass (1977).

3. Stewart and Kamins (1993); Smith (2008); Bulmer, Sturgis, and Allum (2009).

4. Di Gropello (2006).

5. Cooper (2008).