Qualitative modes of data analysis provide ways of discerning, examining, comparing and contrasting, and interpreting meaningful patterns or themes. Meaningfulness is determined by the particular goals and objectives of the project at hand: the same data can be analyzed and synthesized from multiple angles depending on the particular research or evaluation questions being addressed. The varieties of approaches - including ethnography, narrative analysis, discourse analysis, and textual analysis - correspond to different types of data, disciplinary traditions, objectives, and philosophical orientations. However, all share several common characteristics that distinguish them from quantitative analytic approaches.
In quantitative analysis, numbers and what they stand for are the material of analysis. By contrast, qualitative analysis deals in words and is guided by fewer universal rules and standardized procedures than statistical analysis.
We have few agreed-on canons for qualitative data analysis, in the sense of shared ground rules for drawing conclusions and verifying their sturdiness (Miles and Huberman, 1984).
This relative lack of standardization is at once a source of versatility and the focus of considerable misunderstanding. That qualitative analysts will not specify uniform procedures to follow in all cases draws critical fire from researchers who question whether analysis can be truly rigorous in the absence of such universal criteria; in fact, these analysts may have helped to invite this criticism by failing to adequately articulate their standards for assessing qualitative analyses, or even denying that such standards are possible. Their stance has fed a fundamentally mistaken but relatively common idea of qualitative analysis as unsystematic, undisciplined, and "purely subjective."
Although distinctly different from quantitative statistical analysis both in procedures and goals, good qualitative analysis is both systematic and intensely disciplined. If not "objective" in the strict positivist sense, qualitative analysis is arguably replicable insofar as others can be "walked through" the analyst's thought processes and assumptions. Timing also works quite differently in qualitative evaluation. Quantitative evaluation is more easily divided into discrete stages of instrument development, data collection, data processing, and data analysis. By contrast, in qualitative evaluation, data collection and data analysis are not temporally discrete stages: as soon as the first pieces of data are collected, the evaluator begins the process of making sense of the information. Moreover, the different processes involved in qualitative analysis also overlap in time. Part of what distinguishes qualitative analysis is a loop-like pattern of multiple rounds of revisiting the data as additional questions emerge, new connections are unearthed, and more complex formulations develop along with a deepening understanding of the material. Qualitative analysis is fundamentally an iterative set of processes.
At the simplest level, qualitative analysis involves examining the assembled relevant data to determine how they answer the evaluation question(s) at hand. However, the data are apt to be in formats that are unusual for quantitative evaluators, thereby complicating this task. In quantitative analysis of survey results, for example, frequency distributions of responses to specific items on a questionnaire often structure the discussion and analysis of findings. By contrast, qualitative data most often occur in more embedded and less easily reducible or distillable forms than quantitative data. For example, a relevant "piece" of qualitative data might be interspersed with portions of an interview transcript, multiple excerpts from a set of field notes, or a comment or cluster of comments from a focus group.
Throughout the course of qualitative analysis, the analyst should be asking and reasking the following questions:
Two basic forms of qualitative analysis, essentially the same in their underlying logic, will be discussed: intra-case analysis and cross-case analysis. A case may be differently defined for different analytic purposes. Depending on the situation, a case could be a single individual, a focus group session, or a program site (Berkowitz, 1996). In terms of the hypothetical project described in Chapter 2, a case will be a single campus. Intra-case analysis will examine a single project site, and cross-case analysis will systematically compare and contrast the eight campuses.
Qualitative analysts are justifiably wary of creating an unduly reductionistic or mechanistic picture of an undeniably complex, iterative set of processes. Nonetheless, evaluators have identified a few basic commonalities in the process of making sense of qualitative data. In this chapter we have adopted the framework developed by Miles and Huberman (1994) to describe the major phases of data analysis: data reduction, data display, and conclusion drawing and verification.
First, the mass of data has to be organized and somehow meaningfully reduced or reconfigured. Miles and Huberman (1994) describe this first of their three elements of qualitative data analysis as data reduction. "Data reduction refers to the process of selecting, focusing, simplifying, abstracting, and transforming the data that appear in written up field notes or transcriptions." Not only do the data need to be condensed for the sake of manageability, they also have to be transformed so they can be made intelligible in terms of the issues being addressed.
Data reduction often forces choices about which aspects of the assembled data should be emphasized, minimized, or set aside completely for the purposes of the project at hand. Beginners often fail to understand that even at this stage, the data do not speak for themselves. A common mistake many people make in quantitative as well as qualitative analysis, in a vain effort to remain "perfectly objective," is to present a large volume of unassimilated and uncategorized data for the reader's consumption.
In qualitative analysis, the analyst decides which data are to be singled out for description according to principles of selectivity. This usually involves some combination of deductive and inductive analysis. While initial categorizations are shaped by preestablished study questions, the qualitative analyst should remain open to inducing new meanings from the data available.
In evaluation, such as the hypothetical evaluation project in this handbook, data reduction should be guided primarily by the need to address the salient evaluation question(s). This selective winnowing is difficult, both because qualitative data can be very rich, and because the person who analyzes the data also often played a direct, personal role in collecting them. The words that make up qualitative analysis represent real people, places, and events far more concretely than the numbers in quantitative data sets, a reality that can make cutting any of it quite painful. But the acid test has to be the relevance of the particular data for answering particular questions. For example, a formative evaluation question for the hypothetical study might be whether the presentations were suitable for all participants. Focus group participants may have had a number of interesting things to say about the presentations, but remarks that only tangentially relate to the issue of suitability may have to be bracketed or ignored. Similarly, a participant’s comments on his department chair that are unrelated to issues of program implementation or impact, however fascinating, should not be incorporated into the final report. The approach to data reduction is the same for intra-case and cross-case analysis.
With the hypothetical project of Chapter 2 in mind, it is illustrative to consider ways of reducing data collected to address the question "what did participating faculty do to share knowledge with nonparticipating faculty?" The first step in an intra-case analysis of the issue is to examine all the relevant data sources to extract a description of what they say about the sharing of knowledge between participating and nonparticipating faculty on the one campus. Included might be information from focus groups, observations, and indepth interviews of key informants, such as the department chair. The most salient portions of the data are likely to be concentrated in certain sections of the focus group transcripts (or write-ups) and indepth interviews with the department chair. However, it is best to also quickly peruse all notes for relevant data that may be scattered throughout.
In initiating the process of data reduction, the focus is on distilling what the different respondent groups suggested about the activities used to share knowledge between faculty who participated in the project and those who did not. How does what the participating faculty say compare to what the nonparticipating faculty and the department chair report about knowledge sharing and adoption of new practices? In setting out these differences and similarities, it is important not to so "flatten" or reduce the data that they sound like close-ended survey responses. The tendency to treat qualitative data in this manner is not uncommon among analysts trained in quantitative approaches. Not surprisingly, the result is to make qualitative analysis look like watered down survey research with a tiny sample size. Approaching qualitative analysis in this fashion unfairly and unnecessarily dilutes the richness of the data and, thus, inadvertently undermines one of the greatest strengths of the qualitative approach.
Answering the question about knowledge sharing in a truly qualitative way should go beyond enumerating a list of knowledge-sharing activities to also probe the respondents' assessments of the relative effectiveness of these activities, as well as their reasons for believing some more effective than others. Apart from exploring the specific content of the respondents' views, it is also a good idea to take note of the relative frequency with which different issues are raised, as well as the intensity with which they are expressed.
Data display is the second element or level in Miles and Huberman's (1994) model of qualitative data analysis. Data display goes a step beyond data reduction to provide "an organized, compressed assembly of information that permits conclusion drawing..." A display can be an extended piece of text or a diagram, chart, or matrix that provides a new way of arranging and thinking about the more textually embedded data. Data displays, whether in word or diagrammatic form, allow the analyst to extrapolate from the data enough to begin to discern systematic patterns and interrelationships. At the display stage, additional, higher order categories or themes may emerge from the data that go beyond those first discovered during the initial process of data reduction.
From the perspective of program evaluation, data display can be extremely helpful in identifying why a system (e.g., a given program or project) is or is not working well and what might be done to change it. The overarching issue of why some projects work better or are more successful than others almost always drives the analytic process in any evaluation. In our hypothetical evaluation example, faculty from all eight campuses come together at the central campus to attend workshops. In that respect, all participants are exposed to the identical program. However, implementation of teaching techniques presented at the workshop will most likely vary from campus to campus based on factors such as the participants’ personal characteristics, the differing demographics of the student bodies, and differences in the university and departmental characteristics (e.g., size of the student body, organization of preservice courses, department chair’s support of the program goals, departmental receptivity to change and innovation). The qualitative analyst will need to discern patterns of interrelationships to suggest why the project promoted more change on some campuses than on others.
One technique for displaying narrative data is to develop a series of flow charts that map out any critical paths, decision points, and supporting evidence that emerge from establishing the data for a single site. After the first flow chart has been developed, the process can be repeated for all remaining sites. Analysts may (1) use the data from subsequent sites to modify the original flow chart; (2) prepare an independent flow chart for each site; and/or (3) prepare a single flow chart for some events (if most sites adopted a generic approach) and multiple flow charts for others. Examination of the data display across the eight campuses might produce a finding that implementation proceeded more quickly and effectively on those campuses where the department chair was highly supportive of trying new approaches to teaching but was stymied and delayed when department chairs had misgivings about making changes to a tried-and-true system.
Data display for intra-case analysis. Exhibit 10 presents a data display matrix for analyzing patterns of response concerning perceptions and assessments of knowledge-sharing activities for one campus. We have assumed that three respondent units - participating faculty, nonparticipating faculty, and department chairs - have been asked similar questions. Looking at column (a), it is interesting that the three respondent groups were not in total agreement even on which activities they named. Only the participants considered e-mail a means of sharing what they had learned in the program with their colleagues. The nonparticipant colleagues apparently viewed the situation differently, because they did not include e-mail in their list. The department chair - perhaps because she was unaware they were taking place - did not mention e-mail or informal interchanges as knowledge-sharing activities.
Column (b) shows which activities each group considered most effective as a way of sharing knowledge, in order of perceived importance; column (c) summarizes the respondents' reasons for regarding those particular activities as most effective. Looking down column (b), we can see that there is some overlap across groups - for example, both the participants and the department chair believed structured seminars were the most effective knowledge-sharing activity. Nonparticipants saw the structured seminars as better than lunchtime meetings, but not as effective as informal interchanges.
| Exhibit 10. Data matrix for Campus A: What was done to share knowledge |
|||
|
|
Activities named |
Which most effective |
Why |
| Participants |
|
|
|
| Nonparticipants |
|
|
|
| Department chair |
|
|
|
Simply knowing what each set of respondents considered most effective, without knowing why, would leave out an important piece of the analytic puzzle. It would rob the qualitative analyst of the chance to probe potentially meaningful variations in underlying conceptions of what defines effectiveness in an educational exchange. For example, even though both participating faculty and the department chair agreed on the structured seminars as the most effective knowledge-sharing activity, they gave somewhat different reasons for making this claim. The participants saw the seminars as the most effective way of communicating a lot of information concisely. The department chair used indirect indicators - attendance rates of nonparticipants at the seminars, as well as favorable comments on the seminars volunteered to her - to formulate her judgment of effectiveness. It is important to recognize the different bases on which the respondents reached the same conclusions.
Several points concerning qualitative analysis emerge from this relatively straightforward and preliminary exercise. First, a pattern of cross-group differences can be discerned even before we analyze the responses concerning the activities regarded as most effective, and why. The open-ended format of the question allowed each group to give its own definition of "knowledge-sharing activities." The point of the analysis is not primarily to determine which activities were used and how often; if that were the major purpose of asking this question, there would be far more efficient ways (e.g., a checklist or rating scale) to find the answer. From an analytic perspective, it is more important to begin to uncover relevant group differences in perceptions.
Differences in reasons for considering one activity more effective than another might also point to different conceptions of the primary goals of the knowledge-sharing activities. Some of these variations might be attributed to the fact that the respondent groups occupy different structural positions in life and different roles in this specific situation. While both participating and nonparticipating faculty teach in the same department, in this situation the participating faculty are playing a teaching role vis-a-vis their colleagues. The data in column (c) indicate the participants see their main goal as imparting a great deal of information as concisely as possible. By contrast, the nonparticipants - in the role of students - believe they assimilate the material better when presented with smaller quantities of information in informal settings. Their different approaches to the question might reflect different perceptions based on this temporary rearrangement in their roles. The department chair occupies a different structural position in the university than either the participating or nonparticipating faculty. She may be too removed from day-to-day exchanges among the faculty to see much of what is happening on this more informal level. By the same token, her removal from the grassroots might give her a broader perspective on the subject.
Data display in cross-case analysis. The principles applied in analyzing across cases essentially parallel those employed in the intra-case analysis. Exhibit 11 shows an example of a hypothetical data display matrix that might be used for analysis of program participants’ responses to the knowledge-sharing question across all eight campuses. Looking down column (a), one sees differences in the number and variety of knowledge-sharing activities named by participating faculty at the eight schools. Brown bag lunches, department newsletters, workshops, and dissemination of written (hard-copy) materials have been added to the list, which for branch campus A included only structured seminars, e-mail, informal interchanges, and lunchtime meetings. This expanded list probably encompasses most, if not all, such activities at the eight campuses. In addition, where applicable, we have indicated whether the nonparticipating faculty involvement in the activity was compulsory or voluntary.
In Exhibit 11, we are comparing the same group on different campuses, rather than different groups on the same campus, as in Exhibit 10. Column (b) reveals some overlap across participants in which activities were considered most effective: structured seminars were named by participants at campuses A and C, brown bag lunches
| Exhibit 11. Participants’ views of information sharing at eight campuses |
|||
|
|
Activities named |
Which most effective |
Why |
|
|
|
Structured seminar
|
Concise way of communicating a lot of information |
|
|
|
Brown bags | Most interactive |
|
|
|
Structured seminar | Compulsory
Structured format works well |
|
|
|
Combination of the two | Dissemination important but not enough without "personal touch" |
|
|
|
Workshops | Voluntary hands-on approach works best |
|
|
|
Dissemination of materials | Not everyone regularly uses e-mail
Compulsory workshops resisted as coercive |
|
|
|
Lunch meetings | Best time |
|
|
|
Brown bags | Relaxed environment |
by those at campuses B and H. However, as in Exhibit 10, the primary reasons for naming these activities were not always the same. Brown bag lunches were deemed most effective because of their interactive nature (campus B) and the relaxed environment in which they took place (campus H), both suggesting a preference for less formal learning situations. However, while campus A participants judged voluntary structured seminars the most effective way to communicate a great deal of information, campus C participants also liked that the structured seminars on their campus were compulsory. Participants at both campuses appear to favor structure, but may part company on whether requiring attendance is a good idea. The voluntary/compulsory distinction was added to illustrate different aspects of effective knowledge sharing that might prove analytically relevant.
It would also be worthwhile to examine the reasons participants gave for deeming one activity more effective than another, regardless of the activity. Data in column (c) show a tendency for participants on campuses B, D, E, F, and H to prefer voluntary, informal, hands-on, personal approaches. By contrast, those from campuses A and C seemed to favor more structure (although they may disagree on voluntary versus compulsory approaches). The answer supplied for campus G ("best time") is ambiguous and requires returning to the transcripts to see if more material can be found to clarify this response.
To have included all the knowledge-sharing information from four different respondent groups on all eight campuses in a single matrix would have been quite complicated. Therefore, for clarity's sake, we present only the participating faculty responses. However, to complete the cross-case analysis of this evaluation question, the same procedure should be followed - if not in matrix format, then conceptually - for nonparticipating faculty and department chairpersons. For each group, the analysis would be modeled on the above example. It would be aimed at identifying important similarities and differences in what the respondents said or observed and exploring the possible bases for these patterns at different campuses. Much of qualitative analysis, whether intra-case or cross-case, is structured by what Glaser and Strauss (1967) called the "method of constant comparison," an intellectually disciplined process of comparing and contrasting across instances to establish significant patterns, then further questioning and refinement of these patterns as part of an ongoing analytic process.
This activity is the third element of qualitative analysis. Conclusion drawing involves stepping back to consider what the analyzed data mean and to assess their implications for the questions at hand.6 Verification, integrally linked to conclusion drawing, entails revisiting the data as many times as necessary to cross-check or verify these emergent conclusions. "The meanings emerging from the data have to be tested for their plausibility, their sturdiness, their ‘confirmability’ - that is, their validity" (Miles and Huberman, 1994, p. 11). Validity means something different in this context than in quantitative evaluation, where it is a technical term that refers quite specifically to whether a given construct measures what it purports to measure. Here validity encompasses a much broader concern for whether the conclusions being drawn from the data are credible, defensible, warranted, and able to withstand alternative explanations.
6 When qualitative data are used as a precursor to the design/development of quantitative instruments, this step may be postponed. Reducing the data and looking for relationships will provide adequate information for developing other instruments.
For many qualitative evaluators, it is above all this third phase that gives qualitative analysis its special appeal. At the same time, it is probably also the facet that quantitative evaluators and others steeped in traditional quantitative techniques find most disquieting. Once qualitative analysts begin to move beyond cautious analysis of the factual data, the critics ask, what is to guarantee that they are not engaging in purely speculative flights of fancy? Indeed, their concerns are not entirely unfounded. If the unprocessed "data heap" is the result of not taking responsibility for shaping the "story line" of the analysis, the opposite tendency is to take conclusion drawing well beyond what the data reasonably warrant or to prematurely leap to conclusions and draw implications without giving the data proper scrutiny.
The question about knowledge sharing provides a good example. The underlying expectation, or hope, is for a diffusion effort, wherein participating faculty stimulate innovation in teaching mathematics among their colleagues. A cross-case finding might be that participating faculty at three of the eight campuses made active, ongoing efforts to share their new knowledge with their colleagues in a variety of formal and informal settings. At two other campuses, initial efforts at sharing started strong but soon fizzled out and were not continued. In the remaining three cases, one or two faculty participants shared bits and pieces of what they had learned with a few selected colleagues on an ad hoc basis, but otherwise took no steps to diffuse their new knowledge and skills more broadly.
Taking these findings at face value might lead one to conclude that the project had largely failed in encouraging diffusion of new pedagogical knowledge and skills to nonparticipating faculty. After all, such sharing occurred in the desired fashion at only three of the eight campuses. However, before jumping ahead to conclude that the project was disappointing in this respect, or to generalize beyond this case to other similar efforts at spreading pedagogic innovations among faculty, it is vital to examine more closely the likely reasons why sharing among participating and nonparticipating faculty occurred, and where and how it did. The analysts would first look for factors distinguishing the three campuses where ongoing organized efforts at sharing did occur from those where such efforts were either not sustained or occurred in largely piecemeal fashion. However, it will also be important to differentiate among the less successful sites to tease out factors related both to the extent of sharing and the degree to which activities were sustained.
One possible hypothesis would be that successfully sustaining organized efforts at sharing on an ongoing basis requires structural supports at the departmental level and/or conducive environmental conditions at the home campus. In the absence of these supports, a great burst of energy and enthusiasm at the beginning of the academic year will quickly give way under the pressure of the myriad demands, as happened for the second group of two campuses. Similarly, under most circumstances, the individual good will of one or two participating faculty on a campus will in itself be insufficient to generate the type and level of exchange that would make a difference to the nonparticipating faculty (the third set of campuses).
At the three "successful" sites, for example, faculty schedules may allow regularly scheduled common periods for colleagues to share ideas and information. In addition, participation in such events might be encouraged by the department chair, and possibly even considered as a factor in making promotion and tenure decisions. The department might also contribute a few dollars for refreshments in order to promote a more informal, relaxed atmosphere at these activities. In other words, at the campuses where sharing occurred as desired, conditions were conducive in one or more ways: a new time slot did not have to be carved out of already crowded faculty schedules, the department chair did more than simply pay "lip service" to the importance of sharing (faculty are usually quite astute at picking up on what really matters in departmental culture), and efforts were made to create a relaxed ambiance for transfer of knowledge.
At some of the other campuses, structural conditions might not be conducive, in that classes are taught continuously from 8 a.m. through 8 p.m., with faculty coming and going at different times and on alternating days. At another campus, scheduling might not present so great a hurdle. However, the department chair may be so busy that despite philosophic agreement with the importance of diffusing the newly learned skills, she can do little to actively encourage sharing among participating and nonparticipating faculty. In this case, it is not structural conditions or lukewarm support so much as competing priorities and the department chair's failure to act concretely on her commitment that stood in the way. By contrast, at another campus, the department chairperson may publicly acknowledge the goals of the project but really believe it a waste of time and resources. His failure to support sharing activities among his faculty stems from more deeply rooted misgivings about the value and viability of the project. This distinction might not seem to matter, given that the outcome was the same on both campuses (sharing did not occur as desired). However, from the perspective of an evaluation researcher, whether the department chair believes in the project could make a major difference to what would have to be done to change the outcome.
We have begun to develop a reasonably coherent explanation for the cross-site variations in the degree and nature of sharing taking place between participating and nonparticipating faculty. Arriving at this point required stepping back and systematically examining and re-examining the data, using a variety of what Miles and Huberman (1994, pp. 245-262) call "tactics for generating meaning." They describe 13 such tactics, including noting patterns and themes, clustering cases, making contrasts and comparisons, partitioning variables, and subsuming particulars in the general. Qualitative analysts typically employ some or all of these, simultaneously and iteratively, in drawing conclusions.
One factor that can impede conclusion drawing in evaluation studies is that the theoretical or logical assumptions underlying the research are often left unstated. In this example, as discussed above, these are assumptions or expectations about knowledge sharing and diffusion of innovative practices from participating to non-participating faculty, and, by extension, to their students. For the analyst to be in a position to take advantage of conclusion-drawing opportunities, he or she must be able to recognize and address these assumptions, which are often only implicit in the evaluation questions. Toward that end, it may be helpful to explicitly spell out a "logic model" or set of assumptions as to how the program is expected to achieve its desired outcome(s.) Recognizing these assumptions becomes even more important when there is a need or desire to place the findings from a single evaluation into wider comparative context vis-a-vis other program evaluations.
Once having created an apparently credible explanation for variations in the extent and kind of sharing that occurs between participating and nonparticipating faculty across the eight campuses, how can the analyst verify the validity - or truth value - of this interpretation of the data? Miles and Huberman (1994, pp. 262-277) outline 13 tactics for testing or confirming findings, all of which address the need to build systematic "safeguards against self-delusion" (p. 265) into the process of analysis. We will discuss only a few of these, which have particular relevance for the example at hand and emphasize critical contrasts between quantitative and qualitative analytic approaches. However, two points are very important to stress at the outset: several of the most important safeguards on validity - such as using multiple sources and modes of evidence - must be built into the design from the beginning; and the analytic objective is to create a plausible, empirically grounded account that is maximally responsive to the evaluation questions at hand. As the authors note: "You are not looking for one account, forsaking all others, but for the best of several alternative accounts" (p. 274).
One issue of analytic validity that often arises concerns the need to weigh evidence drawn from multiple sources and based on different data collection modes, such as self-reported interview responses and observational data. Triangulation of data sources and modes is critical, but the results may not necessarily corroborate one another, and may even conflict. For example, another of the summative evaluation questions proposed in Chapter 2 concerns the extent to which nonparticipating faculty adopt new concepts and practices in their teaching. Answering this question relies on a combination of observations, self-reported data from participant focus groups, and indepth interviews with department chairs and nonparticipating faculty. In this case, there is a possibility that the observational data might be at odds with the self-reported data from one or more of the respondent groups. For example, when interviewed, the vast majority of nonparticipating faculty might say, and really believe, that they are applying project-related innovative principles in their teaching. However, the observers may see very little behavioral evidence that these principles are actually influencing teaching practices in these faculty members' classrooms. It would be easy to brush off this finding by concluding that the nonparticipants are saving face by parroting what they believe they are expected to say about their teaching. But there are other, more analytically interesting, possibilities. Perhaps the nonparticipants have an incomplete understanding of these principles, or they were not adequately trained in how to translate them effectively into classroom practice.
The important point is that analyzing across multiple group perspectives and different types of data is not a simple matter of deciding who is right or which data are most accurate. Weighing the evidence is a more subtle and delicate matter of hearing each group's viewpoint, while still recognizing that any single perspective is partial and relative to the respondent's experiences and social position. Moreover, as noted above, respondents' perceptions are no more or less real than observations. In fact, discrepancies between self-reported and observational data may reveal profitable topics or areas for further analysis. It is the analyst's job to weave the various voices and sources together in a narrative that responds to the relevant evaluation question(s). The more artfully this is done, the simpler, more natural it appears to the reader. To go to the trouble to collect various types of data and listen to different voices, only to pound the information into a flattened picture, is to do a real disservice to qualitative analysis. However, if there is a reason to believe that some of the data are stronger than others (some of the respondents are highly knowledgeable on the subject, while others are not), it is appropriate to give these responses greater weight in the analysis.
Qualitative analysts should also be alert to patterns of inter-connection in their data that differ from what might have been expected. Miles and Huberman define these as "following up surprises" (1994, p. 270). For instance, at one campus, systematically comparing participating and nonparticipating faculty responses to the question about knowledge-sharing activities (see Exhibit 10) might reveal few apparent cross-group differences. However, closer examination of the two sets of transcripts might show meaningful differences in perceptions dividing along other, less expected lines. For purposes of this evaluation, it was tacitly assumed that the relevant distinctions between faculty would most likely be between those who had and had not participated in the project. However, both groups also share a history as faculty in the same department. Therefore, other factors - such as prior personal ties - might have overridden the participant/nonparticipant faculty distinction. One strength of qualitative analysis is its potential to discover and manipulate these kinds of unexpected patterns, which can often be very informative. To do this requires an ability to listen for, and be receptive to, surprises.
Unlike quantitative researchers, who need to explain away deviant or exceptional cases, qualitative analysts are also usually delighted when they encounter twists in their data that present fresh analytic insights or challenges. Miles and Huberman (1994, pp. 269, 270) talk about "checking the meaning of outliers" and "using extreme cases." In qualitative analysis deviant instances or cases that do not appear to fit the pattern or trend are not treated as outliers, as they would be in statistical, probability-based analysis. Rather, deviant or exceptional cases should be taken as a challenge to further elaboration and verification of an evolving conclusion. For example, if the department chair strongly supports the project's aims and goals for all successful projects but one, perhaps another set of factors is fulfilling the same function(s) at the "deviant" site. Identifying those factors will, in turn, help to clarify more precisely what it is about strong leadership and belief in a project that makes a difference. Or, to elaborate on another extended example, suppose at one campus where structural conditions are not conducive to sharing between participating and nonparticipating faculty, such sharing is occurring nonetheless, spearheaded by one very committed participating faculty member. This example might suggest that a highly committed individual who is a natural leader among his faculty peers is able to overcome the structural constraints to sharing. In a sense, this "deviant" case analysis would strengthen the general conclusion by showing that it takes exceptional circumstances to override the constraints of the situation.
Elsewhere in this handbook, we noted that summative and formative evaluations are often linked by the premise that variations in project implementation will, in turn, effect differences in project outcomes. In the hypothetical example presented in this handbook, all participants were exposed to the same activities on the central campus, eliminating the possibility of analyzing the effects of differences in implementation features. However, using a different model and comparing implementation and outcomes at three different universities, with three campuses participating per university, would give some idea of what such an analysis might look like.
A display matrix for a cross-site evaluation of this type is given in Exhibit 12. The upper portion of the matrix shows how the three campuses varied in key implementation features. The bottom portion summarizes outcomes at each campus. While we would not necessarily expect a one-to-one relationship, the matrix loosely pairs implementation features with outcomes with which they might be associated. For example, workshop staffing and delivery are paired with knowledge-sharing activities, accuracy of workshop content with curricular change. However, there is nothing to preclude looking for a relationship between use of appropriate techniques in the workshops (formative) and curricular changes on the campuses (summative). Use of the matrix would essentially guide the analysis along the same lines as in the examples provided earlier.
| Exhibit 12. Matrix of cross-case analysis linking implementation and outcome factors |
|||||
|
|
|||||
|
|
|
up to date? |
|
|
|
| Campus A |
|
|
|
|
|
| Campus B |
|
|
|
|
|
| Campus C |
|
|
|
|
|
|
|
|||||
|
|
|
|
|
|
|
| Campus A |
|
|
|
|
|
| Campus B |
|
|
|
|
|
| Campus C |
|
|
|
|
|
In this cross-site analysis, the overarching question would address the similarities and differences across these three sites - in terms of project implementation, outcomes, and the connection between them - and investigate the bases of these differences. Was one of the projects discernibly more successful than others, either overall or in particular areas - and if so, what factors or configurations of factors seem to have contributed to these successes? The analysis would then continue through multiple iterations until a satisfactory resolution is achieved.
Issues surrounding the value and uses of conclusion drawing and verification in qualitative analysis take us back to larger questions raised at the outset about how to judge the validity and quality of qualitative research. A lively debate rages on these and related issues. It goes beyond the scope of this chapter to enter this discussion in any depth, but it is worthwhile to summarize emerging areas of agreement.
First, although stated in different ways, there is broad consensus concerning the qualitative analyst's need to be self-aware, honest, and reflective about the analytic process. Analysis is not just the end product, it is also the repertoire of processes used to arrive at that particular place. In qualitative analysis, it is not necessary or even desirable that anyone else who did a similar study should find exactly the same thing or interpret his or her findings in precisely the same way. However, once the notion of analysis as a set of uniform, impersonal, universally applicable procedures is set aside, qualitative analysts are obliged to describe and discuss how they did their work in ways that are, at the very least, accessible to other researchers. Open and honest presentation of analytic processes provides an important check on an individual analyst’s tendencies to get carried away, allowing others to judge for themselves whether the analysis and interpretation are credible in light of the data.
Second, qualitative analysis, as all of qualitative research, is in some ways craftsmanship (Kvale, 1995). There is such a thing as poorly crafted or bad qualitative analysis, and despite their reluctance to issue universal criteria, seasoned qualitative researchers of different bents can still usually agree when they see an example of it. Analysts should be judged partly in terms of how skillfully, artfully, and persuasively they craft an argument or tell a story. Does the analysis flow well and make sense in relation to the study's objectives and the data that were presented? Is the story line clear and convincing? Is the analysis interesting, informative, provocative? Does the analyst explain how and why she or he drew certain conclusions, or on what bases she or he excluded other possible interpretations? These are the kinds of questions that can and should be asked in judging the quality of qualitative analyses. In evaluation studies, analysts are often called upon to move from conclusions to recommendations for improving programs and policies. The recommendations should fit with the findings and with the analysts’ understanding of the context or milieu of the study. It is often useful to bring in stakeholders at the point of "translating" analytic conclusions to implications for action.
As should by now be obvious, it is truly a mistake to imagine that qualitative analysis is easy or can be done by untrained novices. As Patton (1990) comments:
Applying guidelines requires judgment and creativity. Because each qualitative study is unique, the analytical approach used will be unique. Because qualitative inquiry depends, at every stage, on the skills, training, insights, and capabilities of the researcher, qualitative analysis ultimately depends on the analytical intellect and style of the analyst. The human factor is the greatest strength and the fundamental weakness of qualitative inquiry and analysis.
Start the analysis right away and keep a running account of it in your notes: It cannot be overstressed that analysis should begin almost in tandem with data collection, and that it is an iterative set of processes that continues over the course of the field work and beyond. It is generally helpful for field notes or focus group or interview summaries to include a section containing comments, tentative interpretations, or emerging hypotheses. These may eventually be overturned or rejected, and will almost certainly be refined as more data are collected. But they provide an important account of the unfolding analysis and the internal dialogue that accompanied the process.
Involve more than one person: Two heads are better than one, and three may be better still. Qualitative analysis need not, and in many cases probably should not, be a solitary process. It is wise to bring more than one person into the analytic process to serve as a cross-check, sounding board, and source of new ideas and cross-fertilization. It is best if all analysts know something about qualitative analysis as well as the substantive issues involved. If it is impossible or impractical for a second or third person to play a central role, his or her skills may still be tapped in a more limited way. For instance, someone might review only certain portions of a set of transcripts.
Leave enough time and money for analysis and writing: Analyzing and writing up qualitative data almost always takes more time, thought, and effort than anticipated. A budget that assumes a week of analysis time and a week of writing for a project that takes a year’s worth of field work is highly unrealistic. Along with revealing a lack of understanding of the nature of qualitative analysis, failing to build in enough time and money to complete this process adequately is probably the major reason why evaluation reports that include qualitative data can disappoint.
Be selective when using computer software packages in qualitative analysis: A great proliferation of software packages that can be used to aid analysis of qualitative data has been developed in recent years. Most of these packages were reviewed by Weitzman and Miles (1995), who grouped them into six types: word processors, word retrievers, textbase managers, code-and-retrieve programs, code-based theory builders, and conceptual network builders. All have strengths and weaknesses. Weitzman and Miles suggested that when selecting a given package, researchers should think about the amount, types, and sources of data to be analyzed and the types of analyses that will be performed.
Two caveats are in order. First, computer software packages for qualitative data analysis essentially aid in the manipulation of relevant segments of text. While helpful in marking, coding, and moving data segments more quickly and efficiently than can be done manually, the software cannot determine meaningful categories for coding and analysis or define salient themes or factors. In qualitative analysis, as seen above, concepts must take precedence over mechanics: the analytic underpinnings of the procedures must still be supplied by the analyst. Software packages cannot and should not be used as a way of evading the hard intellectual labor of qualitative analysis. Second, since it takes time and resources to become adept in utilizing a given software package and learning its peculiarities, researchers may want to consider whether the scope of their project, or their ongoing needs, truly warrant the investment.
Berkowitz, S. (1996). Using Qualitative and Mixed Method Approaches. Chapter 4 in Needs Assessment: A Creative and Practical Guide for Social Scientists, R. Reviere, S. Berkowitz, C.C. Carter, and C. Graves-Ferguson, Eds. Washington, DC: Taylor & Francis.
Glaser, B., and Strauss, A. (1967). The Discovery of Grounded Theory. Chicago: Aldine.
Kvale, S. (1995). The Social Construction of Validity. Qualitative Inquiry, (1):19-40.
Miles, M.B., and Huberman, A.M. (1984). Qualitative Data Analysis, 16. Newbury Park, CA: Sage.
Miles, M.B, and Huberman, A.M. (1994). Qualitative Data Analysis, 2nd Ed., p. 10-12. Newbury Park, CA: Sage.
Patton, M.Q. (1990). Qualitative Evaluation and Research Methods, 2nd Ed. Newbury Park: CA, Sage.
Weitzman, E.A., and Miles, M.B. (1995). A Software Sourcebook: Computer Programs for Qualitative Data Analysis. Thousand Oaks, CA: Sage.
Coffey, A., and Atkinson, P. (1996). Making Sense of Qualitative Data: Complementary Research Strategies. Thousand Oaks, CA: Sage.
Howe, K., and Eisenhart, M. (1990). Standards for Qualitative (and Quantitative) Research: A Prolegomenon. Educational Researcher, 19(4):2-9.
Wolcott, H.F. (1994). Transforming Qualitative Data: Description, Analysis and Interpretation, Thousand Oaks: CA, Sage.