Discussion Post
Saunders, M. N. K., Lewis, P., & Thornhill, A. (2015). Research methods for business students (7th ed.). Essex, England: Pearson Education Unlimited.
Chapter 5, Section 5.8, “Establishing the Quality of the Research Design” (pp. 201–207)
1. The earlier editions of this book drew a parallel between this procedure and a potent quasi-experimental design, labeled a “nonequivalent, dependent variables design” (Cook & Campbell, 1979, p. 118). According to this design, an intervention may have a variety of relevant outcomes. If a study finds all of them as initially predicted, a conclusion may be reached regarding the effects of the intervention. For instance, in public health studies, some outcomes may have been predicted to be affected, whereas other outcomes may have been predicted not to be affected (Rosenbaum, 2002, pp. 210–211). The empirically determined pattern can then be compared with the initially stipulated pattern. In the quasi-experimental design, the pattern matching occurs in the following manner: If, for each outcome, the initially predicted values have been found, and at the same time alternative “patterns” of predicted values (including those deriving from methodological artifacts, or “threats” to validity) have not been found, causal inferences can be entertained.
2. The lack of attention to the transitions has possibly arisen because of the graphic confusion between a logic model and a flowchart. A flowchart, as used in its original applications in industrial engineering, merely indicates that one box is followed by another—analogous to an assembly line. In a logic model, the lines presuppose more than a simple sequential relationship. The lines also represent some kind of triggering process—that one box producesthe next one. How the triggering occurs is then the transition requiring careful explanation when using logic models.
3. Researchers who have previously been heavily engaged in doing quantitative research may struggle with the case-based approach, because variable-based thinking may be a subconscious part of their natural research orientation. Such a speculation may be unfair. However, my consultations with different research teams have suggested that researchers accustomed to doing quantitative studies (but who then plan to do a multiple-case study) readily think in terms of variables as the key elements in any analysis. Such thinking then leads to wanting to do any analysis according to those variables, despite the small number of cases available.
APPLICATION #7: Using a Case Study to Compare Directly Competing Rival Hypotheses: Whether Military Base Closures Produce Catastrophic Economic Impacts or Not
In experimental research, the use of a control group represents an attempt to rule out all rivals—but without specifying or investigating them. Although case study research does not offer the same opportunity, the number of plausible rivals may be small, and investigating them directly still can be manageable. As a result, entertaining and directly examining individual rival hypotheses can markedly strengthen a case study. Application 7 shows how a case study addressed its main proposition and its main rival, indicating how the evidence supported one but not the other.
Military bases located throughout this country not only fulfill important military functions but also can make valuable contributions to local economies. By employing portions of the local civilian population and consuming local resources, and especially by being located in small jurisdictions, a military base can play a substantial economic role in a jurisdiction.
When such bases are then closed, usually in relation to the reorganization and consolidation of bases across the country, the closures pose a dire threat to the local economy. Such was the case with an Air Force base in a rural county in California, and the base’s closure was the subject of a case study by Ted K. Bradshaw. 1
1. Bradshaw, T. K. (1999). Communities not fazed: Why military base closures may not be catastrophic. Journal of American Planning Association, 65, 193–206. The present author summarized this article, which then appeared as Chapter 18 in Yin (2004), The Case Study Anthology. Readers should consult the original journal article to appreciate its full scope, covering six sectors and containing supporting tables and graphs. Due to space limitations, these materials are not reproduced in this Application 7.
Hypothesis 1.
The initial hypothesis was that the base closure would have a “catastrophic” impact on the county, for the following reasons.
The base was a well-established Strategic Air Command facility for B-52 bomber and K-135 tanker crews. It employed more than 6,000 persons (5,000 military and 1,000 civilian), making it the county’s largest employer, representing 10% of the county’s employees. Similarly, 11,000 military personnel, spouses, and dependents were associated with the base, representing 6% of the county’s population. Moreover, the county’s broader economy was dominated by agriculture and related industries and did not have other large employers or other federal government facilities to which the base’s 1,000 civilian employees could transfer.
The base closure, following the typical congressional and public objections over such closures, had been one of those recommended by the Base Realignment and Adjustment Commission. As a result, the base’s operations and most of its personnel were transferred to other military installations located in Oklahoma and Louisiana, during the year prior to the formal closing of the base.
At that time, a formal task force report predicted the dire economic consequences that soon would occur. The report said that the county would suffer a loss of 3,700 civilian jobs, a population loss of 18,000 persons, and a loss of $105 million in retail sales. The county’s unemployment rate, already chronically high at 14.4%, was predicted to rise to 21.7%. All these potential job, population, sales, and unemployment levels were interpreted as representing a catastrophic outcome for the county’s economy.
Hypothesis 2.
In support of a contrary hypothesis, the case study started by pointing to the findings from studies of other base closures that had taken place several years earlier.
One of the studies had examined the impacts of closures of three bases in the same state as the Strategic Air Command facility. This study as well as the others all suggested that these base closures had not been accompanied by catastrophic effects, even in the short run. Although some economic decline did occur, the impact was not as severe as had been predicted. Furthermore, in the long run, the abandoned base facilities also provided the opportunity for renewed economic development.
Bradshaw’s case study then examined the two competing rival hypotheses. He collected and presented a variety of quantitative (economic) data before, during, and after the year of the base closure, in six important economic sectors: retail sales, local equipment suppliers, hospital and health care services, employment and unemployment, housing, and population change. In each sector, the case study found that strong negative effects had been avoided, and the main conclusion was that the closure had not produced a catastrophic outcome.
More important were the author’s explanations of why a catastrophic outcome had not occurred, at the same time showing how qualitative data readily complemented the economic data. These explanations were based on Bradshaw’s interviews with key local officials, community and business representatives, and military staff. Illustratively, for the purposes of this Application 7, the experiences in three of the sectors are discussed next, involving retail sales, local equipment suppliers, and employment and unemployment [the original case study presented data for all six sectors—see footnote 1 above].
Explanations for changes in three illustrative economic sectors.
In the retail sales sector, the original fear of a great reduction in sales, loss of retail jobs, and diminished local tax revenue was not realized because much of the base’s retail purchasing had been done at the base’s commissaries, not the county’s local outlets. Except for a number of outlets located near the base, such purchasing power, therefore, had not been part of the local economy in the first place. Moreover, a modestly sized population group associated with the base—military retirees—remained in the community after the base operations had transferred. These retirees then had to shift their purchasing from the (closed) commissaries to the local outlets, thereby creating a small positive impact on the county’s retail sales.
With regard to local equipment suppliers, the base had been undergoing a major construction project that had led to the Air Force’s procuring of local equipment and equipment services. This activity did cease with the base closure, but instead of creating a total void, the remaining base operations involved the initiation of a new program, to clean up the toxic waste left behind by the closed base. Although the original suppliers might not have been involved in the new program, the overall economic effect was more balanced than had been expected.
The employment and unemployment trends had been examined over a 5-year period, bracketing the year of the base closure. Strong seasonal effects had required that year-to-year comparisons be made, focusing on the comparable months from one year to the next. Such comparisons for the month of April, the actual month of the base closure, had indicated a slight increase in the unemployment rate from before the closure to the year of the closure and then a slight decline in the ensuing 2 years. Comparisons for the several Octobers or for Januarys had resulted in similar patterns.
At the same time, a potential difficulty in interpreting the employment and unemployment trends was that the county was in an economically growing state and region. Therefore, a claim could be made that the employment-unemployment picture would have been rosier, rather than roughly neutral, had the base not been closed. Bradshaw explored this possibility by examining data from the neighboring counties. Although their trends were better, the differences were far from dramatic, much less supporting any catastrophic interpretation related to the base closure.
Conclusion.
The author concluded that, although catastrophic effects had not occurred in this case, such effects still could occur in other cases. For instance, the consequences could differ if the military base involved a large manufacturing or research-and-development component that employed many civilian workers. Nevertheless, such conditions did not exist in the present case, and the case study aptly explained the reasons that the catastrophic effects predicted by Hypothesis 1 had not occurred.
FOR CLASS DISCUSSION OR WRITTEN ASSIGNMENT
Thinking About Follow-up Studies
No single-case study (or any other type of research study, for that matter) is likely to present such a final set of findings and conclusions that no further inquiries will be relevant. On the contrary, to complement and augment earlier studies, additional studies should always be welcome. For instance, as your own research study draws to an end, you should be thinking about the ways of doing another study (e.g., to replicate the original findings, to strengthen them, or to extend them to a newer set of issues or even situations).
Application 7 appears to have produced a sound set of findings in support of its major conclusion. Especially notable is that the research drew upon extensive quantitative (economic) data and also involved an array of interviews with relevant local officials, community representatives, and military staff [remember again that much of the original work had to be omitted from this application due to space limitations]. The study also cites complementary findings by other studies of base closures, conducted several years earlier.
Discuss how you might think about doing a follow-up study, expanding the original study design and not just augmenting the array of evidence. One possibility would be to go outside the target community and to conduct another case study in a comparable community where a similar base closure had occurred. Another possibility would be to collect fresh data from the original target community, but at a later date. For instance, residents who might still remain in the community might be asked in a survey to describe their current living and working situations as well as to recollect the earlier events as closure was being announced and taking place. Are there yet other possibilities?
APPLICATION #8: A Nutshell Example of an Explanatory Case Study: How a Federal Award Affected a University Computer Department
Application 8 was not originally a case study but comes from the abstract of a final grant report, submitted by the grant’s principal investigator. 1 The abstract is presented in its original form, with methodological comments [in bold and brackets] added by the present author.
1. This application, with minor edits, originally appeared as Chapter 7 in Yin (2012a), Applications of Case Study Research.
In the abstract, the original author attempts to attribute significant organizational changes in a university computer science department to the use of funds from the federal grant. Because of the abstract’s limited two-page (and originally single-spaced) length, the abstract does not try to present the data or evidence to support its claims. However, the essence of the logic serves as an excellent point of departure for understanding how to frame an explanatory case study, even though the illustration comes from an earlier era in the evolution of academic computer science.
During the past seven years, the Computer Science Department at Cornell was radically transformed from theoretical, pencil-and-paper research operation to one with a high degree of experimental computing. The departmental computing facility grew from a VAX/780 and a PDP11/60 to an integrated complex of almost 100 workstations and UNIX mainframes [the funded initiative]. All faculty and graduate students now use these computers daily [a sequentially earlier outcome, further itemized below], and much research that was hitherto impossible for us is now being performed [a sequentially later outcome, operationalized further below].
The change in emphasis was due to the maturing of computer science [a potential rival explanation, to be examined had there been more space], to commensurate changes in the interests of the faculty [another potential rival explanation], and to hardware and software advances that made flexible computing available at an affordable price [a third rival]. However, without the National Science Foundation’s five-year grant, it would not have been possible [the main hypothesized explanation]. The grant provided the wherewithal that allowed the department to change with the times; it provided equipment and maintenance, gave us leverage with vendors for acquiring other equipment, and funded staffing of the faculty [critical how-and-why enhancement of the main explanation, indicating how the grant worked to produce the outcomes described next].
The influence of the grant can be seen by mentioning just a few of the more important projects that it has stimulated. Turing Award winner John Hopcroft changed his interests [with additional space, the text could have explained how and why the grant led to these changed interests] from the theory of algorithms and computational complexity to robotics and now heads a growing and forceful group that is experimenting with robotics and solid modeling [operational outcome]. Theoretician Robert Constable and his group have been developing a system of “mechanizing” mathematics. This system, which has inspired many theoretical as well as experimental advances, has as one of its goals the extraction of a program from a mathematical proof; it gives a glimpse into how professional programming might be done 20 years from now [a second operationalized outcome]. Tim Teitelbaum and his group generalized his work on the well-known Program Synthesizer into a system that is able to generate such a programming environment from a formal description of a language; the resulting Synthesizer Generator has been released to more than 120 sites worldwide [a third]. Ken Birman’s group is developing an experimental distributed operating system for dealing with fault tolerance [a fourth]. And visitor Paul Pritchard used the facility for his work on prime numbers, resulting in the first known arithmetic progression of 19 primes [a fifth].
The grant enabled the department to attract bright young faculty who would not have joined a department with inadequate facilities [beginning of a broader explanation, suggesting how the grant affected the whole department]. As a result, the department has been able to branch out into new areas, such as VLSI, parallel architectures and code optimization, functional programming, and artificial intelligence [continued explanation]. The grant program did what it set out to do: It made it possible for the department to expand its research activity, making it far more experimental and computing intensive while still maintaining strong theoretical foundations [summary explanation].
FOR CLASS DISCUSSION OR WRITTEN ASSIGNMENT
Relying on Self-Reports
When a document takes the form of the grant report just cited in Application 8, you would consider it to be one type of “self-report.” The author of the document has narrated a particular version of events and ideas. The document is a form of self-report because no references are made to other sources that might corroborate the document’s contents. Another type of self-report arises when you interview a single person and cite her or his rendition of reality without trying to corroborate the information.However, sometimes the document or the interview is the entirety of the reality. In other words, the document’s author or the interviewee is expressing her or his own perspective, opinion, or attitude, without regard to any external reality. Furthermore, your case study might have been highly interested in capturing and examining just that particular point of view, especially if, for instance, you were doing a social justice study. In those situations, what might appear to be limited to being considered a “self-report” might contain revealing and precious insights that by definition are not subject to any corroboration.
Possibly citing some of your own research, clarify and discuss the situations when corroborating self-reports appears important, compared with when it does not appear to be needed, much less wanted.
APPLICATION #9: An Explanatory Case Study: Transforming a Business Firm Through Strategic Planning
Explanatory case studies can examine a complexity of activities and events, such as the transformation of a business firm. Application 9 contains a complete explanatory case study about one firm, defining the breadth of changes covered by the transformation and suggesting how a strategic planning process was instrumental in the transformation.
The business firm in Application 9, Bolt, Inc., was a family-owned machine shop and components manufacturer located in Grand Prairie, Texas. 1 The firm had been pressured by its major customer to improve its production system or risk losing new orders. In response, implementation of cellular manufacturing solved the firm’s initial production problems and resulted in a 300% capacity increase and a rise in employee skill levels and problem-solving capabilities. However, in addition to the improved manufacturing processes, the firm’s management team developed a shared, common direction about what to do with the company’s extra capacity. The team undertook a strategic planning process that set the course for achieving important long-range company goals and objectives in marketing, information systems, manufacturing, and human resources.
1. This application originally appeared as Chapter 9 in Yin (2012a), Applications of Case Study Research, in which the present author composed a condensed and edited version of a complete case study by Jan Youtie. She had worked under the direction of the present author, who designed the original multiple-case study. To conserve space, several exhibits and numerous footnotes (citing specific interview and documentary sources) in the original case study have been omitted.