In his extremely popular and often quoted article (Flyvbjerg, B. (2006). Five misunderstandings about case-study research. Qualitative inquiry, 12(2), 219-245.) Flyvbjerg presents five answers to five misunderstandings about case studies.
The article has been published twice:
Sosiologisk tidsskrift 12 (2), 117-142, 2004 - 1658 citations according to google scholar.
Qualitative Inquiry. Paper listed as the "Most Read" and "Most Cited”, 2006 - 19110 citations according to google scholar.
The article is also used in academic text books on qualitative research:
Creswell, John W., and Cheryl N. Poth. Qualitative inquiry and research design: Choosing among five approaches. Sage publications, 2016. - 152637 citations according to google scholar.
Green, Judith, and Nicki Thorogood. Qualitative methods for health research. sage, 2018. - 7563 citations according to google scholar.
In the danish context of pædagogy and education which concerns the present study, it has been citet in the often used book:
Kvale, Steinar, and Svend Brinkmann. Interview: introduktion til et håndværk. Hans Reitzels Forlag, 2009. - danish version with 1436 citations according to google scholar.
Kvale, Steinar, and Svend Brinkmann. Interviews: Learning the craft of qualitative research interviewing. sage, 2009. - english version 20951 citations according to google scholar.
Suffice it to say that Flyvbjerg’s article is very popular.
The article is cited primarily for two reasons. First, and specifically in Kvale & Brinkmann, as a defense against common misunderstandings and attacks against the validity of case-studies (Kvale, Steinar, and Svend Brinkmann 2009:291).
Secondly the article is cited for the distinctions made by Flyvbjerg between types of cases.
The first reason is discussed here. The great popularity of the article warrants a discussion of the conflicting arguments that are presented in it.
On Flyvbjerg’s arguments
In the article Flyvbjerg chronologically gives answers to five misunderstandings about case-studies, I refer to these answers as answer 1, 2 etc.
Flybjergs answers 1 and 2 are directly opposed to each other. Regarding answer 1 he concludes that there can be no predictive, context-independent theories in his field, which he interchangeably calls “human affairs” and “social sciences”. With regards to the second answer however he concludes that:
“One can often generalize on the basis of a single case, and the case study may be central to scientific development via generalization as supplement or alternative to other methods.” (s12)
While only implying it in the above quote, his argument is that a case-study can be generalized, whenever it falsifies a hypothesis. This is why the section of the text in which this argument is made is called: “Cases as Black Swans”. Toward the end of the section Flyvbjerg partly concludes:
“Popper himself used the now famous example of, ‘All swans are white,’ and proposed that just one observation of a single black swan would falsify this proposition and in this way have general significance and stimulate further investigations and theory-building. The case study is well suited for identifying ‘black swans’ because of its in-depth approach: what appears to be ‘white’ often turns out on closer examination to be ‘black.’” s11
Fittingly all of the examples given of generalizable case studies, are all examples of the falsification of predictive, context-independent theories. This fact however means that Flyvbjergs answer to misunderstanding 1, clashes directly with his answer to misunderstanding 2. If, as is said in the first answer:
“Predictive theories and universals cannot be found in the study of human affairs.” (s7)
Then the reader must make a choice that is not confronted in the text. Either there are predictive theories in “human affairs” (of which Flyvbjerg gives several examples), or case-studies cannot be generalized in that discipline. This problem also involves his answer to the third misunderstanding which is:
“The case study is useful for both generating and testing of hypotheses but is not limited to these research activities alone. “ (s13)
How can we generate and test hypotheses in human affairs, if there are no predictive theories, to which such hypotheses would apply?
Likewise his fourth answer becomes problematic:
“The case study contains no greater bias toward verification of the researcher’s preconceived notions than other methods of inquiry. On the contrary, experience indicates that the case study contains a greater bias toward falsification of preconceived notions than toward verification.” (s21)
How can we claim that case-studies are useful for falsification, if there are no predictive theories, with hypotheses to falsify?
The easy answer to all of these questions, is that Flyvbjerg’s answer 1 is simply wrong. We can absolutely make predictive theories about human affairs. We might have many more difficulties in coming up with good tests to falsify them, than in disciplines concerned with more simple aspects of our world, but it is certainly possible. As noted, Flyvbjerg himself mentions numerous cases throughout his text:
“John Goldthorpe et al. (1968, 1969) deliberately looked for a case that was as favorable as possible to the thesis that the working class, having reached middle-class status, was dissolving into a society without class identity and related conflict (see also Wieviorka 1992)” (s9)
Indeed the marxist idea of classes in modern societies definitely holds a set of testable hypotheses, about expected behaviors and attitudes among groups of people (human affairs), which together form a theory that allows us to make predictions. We cannot be certain of the universality of everything we find, or the precision of the methods that we use, but a predictive theory it most certainly is.
Two more examples:
“A model example of a ‘least likely’ case is Robert Michels’s (1962) classical study of
oligarchy in organizations. By choosing a horizontally structured grassroots organization with strong democratic ideals--that is, a type of organization with an especially low probability of being oligarchical--Michels could test the universality of the oligarchy thesis; that is, ‘If this organization is oligarchic, so are most others.’ A corresponding model example of a ‘most likely’ case is W. F. Whyte’s (1943) study of a Boston slum neighborhood, which according to existing theory should have exhibited social disorganization, but in fact showed quite the opposite (see also the articles on Whyte’s study in Journal of Contemporary Ethnography, vol. 21, no. 1, 1992).” (s15)
There is no reason, according to the examples and arguments that Flyvbjerg makes, to retain the position that there cannot be predictive theories in the study of human affairs. We should reject the response to the first misunderstanding about case studies. This rejection does not however, mark an end to our problems.
What is a case-study?
When Flyvbjerg presents examples of case studies that falsify some hypotheses he initiates with the example of Galileo dropping spheres from the Tower of Pisa (s8-9). While there is some speculation as to whether Galileo actually did this, we may comment that most people would refer to the situation with the word “experiment” rather than the words “case” and “study”. Flyvbjerg mentions a few examples of case studies, which seem at odds with the nature of the kinds of case studies that he supposedly defends:
“The extreme case can be well-suited for getting a point across in an especially dramatic way, which often occurs for well-known case studies such as Freud’s ‘Wolf-Man’ and Foucault’s ‘Panopticon.’” (s14)
It is fair to say that the case of the “Wolf-Man” is a case study. Panopticon on the other hand is definitely not a case study although it did entail a variety of hypotheses to be tested about the behavior of humans under surveillance. While there have been many case studies concerned with Foucaults theoretical framework from The Panopticon Writings, the original text was only a case study in the sense that Foucault wrote in reaction (not as receiver) to letters by Jeremy Bentham. Were we to accept such a broad definition of case study, we should include every theoretical text to draw upon elements from the real world. But who would say that the misunderstandings answered in Flyvbjerg’s text concern the texts of Michel Foucault?
In other words Flyvbjerg stretches the concept of “case studies” too far, and does not give a formal definition. Consequently the reader can only infer what kind of case study the misunderstandings relate to.
While it is only implied what kind of case study Flyvbjerg is actually defending, the general understanding of the concept narrows toward the end of the text, with answer 5:
“It is correct that summarizing case studies is often difficult, especially as concerns case process. It is less correct as regards case outcomes. The problems in summarizing case studies, however, are due more often to the properties of the reality studied than to the case study as a research method. Often it is not desirable to summarize and generalize case studies. Good studies should be read as narratives in their entirety.” (s25)
And here is, for comparison, the first sentence of Flyvbjergs answer to misunderstanding number two: “One can often generalize on the basis of a single case…” (s12)
There is no logical inconsistency between the last sentence of Flyvbjergs fifth answer, and the first sentence of his second answer. Just because something is possible, does not mean that it is desirable. The problem is that Flyvbjerg basically negates the usefulness of his argument against misunderstandings two, three and four. The position laid forth in favor of the generalizability of case studies, becomes miniscule. Case studies can be used to generalize (2), and can therefore be used to falsify hypotheses and theories (3), and therefore have a greater bias toward falsification than verification (4), but we should not use them in that way (5).
“Good studies should be read as narratives in their entirety.” (s25)
A good case study, in other words, is a case study that should neither be summarized or generalized. It should be read in its entirety.
This fits well with Flyvbjerg’s underlying assumptions about research:
“If one thus assumes that the goal of the researcher’s work is to understand and learn about the phenomena being studied, then research is simply a form of learning. If one assumes that research, like other learning processes, can be described by the phenomenology for human learning, it then becomes clear that the most advanced form of understanding is achieved when researchers place themselves within the context being studied. Only in this way can researchers understand the viewpoints and the behavior, which characterizes social actors.” (s20)
This position however weakens the argument for publicly funded research. Surely the end product of research should be to communicate the results so as to improve the lives of people? What good is it to everyone else, that some researchers know alot about human affairs, if our conception of learning is that it is only possible to achieve the same knowledge by actually doing the research oneself?
Returning to Flyvbjerg’s answers it is now clear that they can be grouped. One group consists of answers 1 and 5, which fit together. The other group consists of answers 2, 3 and 4, which do not fit with the first group.
In the first group, answers 1 and 5 present a situation in which case studies are about context dependent-knowledge, they result in condensed narratives, full of detail about the many strange ways of the world. They should seldomly be summarized or generalized. For this reason, any affordances toward predictive theories would be rare. If it falsifies a hypothesis, then it is not hard to summarize.
In the second group, answers 2, 3 and 4 provide us with a different picture. Experiments as well as theoretical discussions can be considered case studies. This also means that case studies can often be generalized and are great for falsifying hypotheses and challenging theories.
Seeing what actually goes on
One of the most popular academic textbooks on the subject of qualitative research, is Case Study Design: Research and Methods by Robert K. Yin, who makes a very important distinction between two types of generalization. He distinguishes between those generalizations that are analytical and those that are statistical (Yin, 2009: 15, 38-39, 43). Yin writes:
“The short answer is that case studies, like experiments, are generalizable to theoretical propositions and not to populations or universes. In this sense, the case study, like the experiment, does not represent a “sample,” and in doing a case study, your goal will be to expand and generalize theories (analytical generalization) and not to enumerate frequencies (statistical generalizations). ” - (Yin, 2009: 15)
The often used metaphor regarding black and white swans is an example of an analytical generalization. We need only one observation to reject hypotheses of the absolute kind. For example if the hypothesis requires that all swans are white. With a small twist however, we can quickly make that hypothesis into a statistical hypothesis:
“Most swans are white” or “there are more black swans than there are white swans”.
In order to reject one of these hypotheses we should collect enough observations of swans of any other color than white. Due to the burdensome nature of field work, we would quickly enter the territory of statistical inference, probability and estimation.
Flyvbjerg does not address this important distinction, which, with some rhetorical adjustments, would lend merit to the claim that: “Predictive theories and universals cannot be found in the study of human affairs.” (s7). The statement is still wrong, but it is a limitation on the affordance of case-studies to predictive theories, that they generally do not make statements about frequencies more plausible. Human affairs are so complex and diverse, that it is rarely possible to speak about them in absolute terms.
Ingen kommentarer:
Send en kommentar