Qualitative Health Research, 1996 (November), 6 (4), 561-567 


Q METHODOLOGY AND QUALITATIVE RESEARCH

Steven R. Brown

Kent State University

Q methodology was invented in 1935 by British physicist-psychologist William Stephenson (1953) and is most often associated with quantitative analysis due to its involvement with factor analysis. Statistical procedures aside, however, what Stephenson was interested in providing was a way to reveal the subjectivity involved in any situation -- e.g., in aesthetic judgment, poetic interpretation, perceptions of organizational role, political attitudes, appraisals of health care, experiences of bereavement, perspectives on life and the cosmos, et cetera ad infinitum. It is life as lived from the standpoint of the person living it that is typically passed over by quantitative procedures, and it is subjectivity in this sense that Q methodology is designed to examine and that frequently engages the attention of the qualitative researcher interested in more than just life measured by the pound. Q methodology "combines the strengths of both qualitative and quantitative research traditions" (Dennis & Goldberg, 1996, p. 104) and in other respects provides a bridge between the two (Sell & Brown, 1984).

Some of the quantitative obstacles to the wider use of Q methodology have recently been rendered less daunting by virtue of software packages which have converted to button presses what before were tedious calculations. One such package, QMethod (Atkinson, 1992), is available as freeware from Kent State University's Listserver, and the way in which QMethod facilitates Q-methodological inquiries, as well as the way in which such inquiries proceed, is perhaps best illustrated in terms of a concrete example such as the following.

During the course of a small-scale study of health care, attention came to be focused on the quality of care received by a particular individual (who will be referred to as K), specifically on his experiences as viewed against the backdrop of prior care. The treatment to which K was recently exposed was of only moderate seriousness, involving preoperative testing, anesthesia, surgery, and a two-day hospital stay.

The instrumental basis of Q methodology is the Q-sort technique, which conventionally involves the rank-ordering of a set of statements from agree to disagree. Usually the statements are taken from interviews, hence are grounded in concrete existence; for purposes of convenience, however, the Q sample in this case consisted of 24 statements taken from Larson's (1984) CARE-Q set. (Q samples can also be composed of pictures, recordings, and any other stimuli amenable to appraisal.) K was initially invited to characterize the care rendered by his surgeon by sorting the 24 statements (each typed on a separate card) into a quasi- normal distribution ranging from "most like the care given by my surgeon" (+4) to "most unlike" (-4), the result being the Q sort shown in Table 1. The Q-sorting session was followed by a focused interview during which K was invited to expand on his experience.

Table 1:
Q Sort Representing Surgeon's Care

-4  -3  -2  -1   0  +1  +2  +3  +4 
---------------------------------- 
 1   3  10   4   5   2  11   6  20 
 9  21  13  18   7  14  12   8  24 
        22  23  15  17  19 
                16

As indicated, K agreed most strongly with statements 20 and 24, and disagreed in equal measure with 1 and 9, which read as follows:

     20.  Was honest with me about my condition 
     24.  Gave me good physical care 

      1.  Volunteered to do "little" things for me 
      9.  Touched me when I needed comforting

As is apparent, the surgeon's care was of a "professional" kind -- competent, informative, direct -- and this was all that K desired and expected.

But there were other expressions of care to which K was exposed during his stay. At one point, for example, Nurse A had to catheterize him, but did so only after explaining the procedure and the need for it. A few short hours later, Nurse B insisted on catheterizing K again for no good reason which he thought balanced the pain involved; his objections brought Nurse C, who overrode B's decision. The care provided by each of these nurses (as experienced by K) was also incorporated into the study, each as a separate Q sort. K also represented the kind of care he remembered receiving at a university health service when he was a student (his only other hospitalization); also represented as a Q sort was the care he remembered receiving at the hands of his mother during various childhood illnesses. All of these Q sorts were rendered by K, each from his own subjective standpoint.

QMethod is a mainframe Fortran program specifically designed for Q-sort data and compatible with Stephenson's broader methodological principles. It begins by facilitating the entry of the statements used in the study and the dimensions of the particular Q sort employed. It then prompts the analyst to enter each Q sort in turn (maximum n=120 sorts). Once the data are entered, the analyst then selects the correlation option (from a menu), followed by the factor-analytic option. The program is quite fast, even with relatively large matrices. Once the original factors have been extracted, the analyst has the option of rotating them either by varimax criteria, or judgmentally according to theoretical considerations. In the latter, the vectors are displayed graphically so that the analyst can visually inspect the effects of the rotations undertaken. The final step (also from a menu) involves calculation of factor scores and the incorporation of all tables in an output file. Some basic knowledge of factor theory is essential for proper use of the QMethod program, but not nearly to the extent previously required; moreover, elementary introductions to all aspects of Q methodology are now available (e.g., Brown, 1980, 1993; McKeown & Thomas, 1988; Stainton Rogers, 1995; Stainton Rogers, 1991, pp. 126-133).

Table 2:
Correlation and Factor Matrices

                      Correlations(a)         Factors(b)
K's Q Sorts        1   2   3   4   5   6      I  II  III
_________________________________________________________

1 Nurse A         ..  24  30  71 -12  55      X
2 Nurse B         24  .. -18  22 -47  28
3 Nurse C         30 -18  ..  34   0  22          X
4 Surgeon         71  22  34  .. -29  64      X
5 Mother         -12 -47   0 -29  .. -10              X
6 University      55  28  22  64 -10  ..      X
_________________________________________________________
 (a) Decimals to two places omitted from correlations.
 (b) X = significant loadings; others insignificant.

Two of the more revealing output tables from QMethod are shown in Table 2. The correlation table shows the extent to which K's different experiences of care were similar or different. For instance, K experienced the surgeon's demeanor as similar to Nurse A's (r = .71) but as unlike his mother's (r = -.29). The factor matrix summarizes which of the Q sorts are similar to or different from one another, i.e., which represent different experiences of care: Hence, the surgeon, Nurse A, and K's university hospitalization were experienced as variations on a common theme (i.e., the statement rankings in these three Q sorts were highly similar); Nurse C, who saved K from an unnecessary catheterization, defined another factor; K's mother defined a third. In sum, K's experiences have included three different kinds of caring.

Table 3:
Factor Arrays

                                                           I  II  III
______________________________________________________________________

20  Was honest with me about my condition                  4   2  -1
12  Was well organized                                     3   1  -3

10  Knew when I had had enough, acted accordingly         -4   4   0
11  Included me in planning, management of my care         1   3  -2

21  Talked to me                                          -1   0   3
 9  Touched me when I needed comforting                   -4  -4   2
______________________________________________________________________

The most important output table from QMethod is the table of factor scores (ranging in this case from +4 to -4), which indicates the extent to which each of the 24 statements characterizes each of the three factors. It is in this table that the main features of K's subjectivity are rendered manifest. As Table 3 shows, factor I (which subsumes the surgeon, Nurse A, and K's university hospitalization) involves care which K experienced as straightforwardly informational and well organized (statements 20, 12; compare with scores in factors II and III). Factor II implicates Nurse C only, and statements 10 and 11 show how she empowered K and returned to him some degree of control over his own care (by overriding Nurse B's earlier decision to catheterize him again). Factor III involves a mother's care -- of comforting K by talking and touching (statements 21, 9). We would not be too wide of the mark were we to label K's factors as Professional (I), Empowering (II), and Compassionate (III), each a subjectively distinguishable experience of care. Needless to say, the three factors are necessarily grounded in K's experiences, from which they have arisen.

This brief article precludes detailed consideration of technicalities and the deeper philosophical issues that render Q methodology unique. Suffice it to say that despite its mathematical substructure, Q's purpose is to reveal subjective structures, attitudes, and perspectives from the standpoint of the person or persons being observed. Those interested in further details can consult the growing literature in the health field (e.g., Dennis, 1986, 1990; Dennis & Goldberg, 1996; Mrtek, Tafesse, & Wigger, in press; Stainton Rogers, 1991; Wigger & Mrtek, 1994).

Persons wishing to obtain a copy of the QMethod program (available in IBM, VAX, and UNIX versions) can send the command "get qmethod package" (IBM), "get vaxq package" (VAX), or "get unixq package" (UNIX) as the sole content of an email note addressed to Listserv@listserv.kent.edu. This will result in a list of GET commands for all files in the package; this list can then be forwarded to the listserv address and the files will be forwarded. Peter Schmolck has modified QMethod for PCs, a copy of which can be obtained from his Web site: http://www.rz.unibw-muenchen.de/~p41bsmk/qmethod/. Michael Stricklin's PCQ program is comparable to QMethod, but is not freeware; Stricklin can be reached at mstrick@unlinfo.unl.edu.

Those interested in finding out more about Q methodology can do so by joining the Q electronic discussion group: Send the command "subscribe Q-Method Chris Jones" (replace Chris Jones with your own name) to the Listserv address above. Various documents and manuscripts are available from the QArchive at the University of Wisconsin, Whitewater (http://facstaff.uww.edu/cottlec/qarchive/qindex.htm). Additional information and applications are also to be found in the Q journal Operant Subjectivity, now in its seventeenth volume (send subscription inquiries to the Stephenson Research Center, School of Journalism, University of Missouri, Columbia, MO 65205), and in presentations at the annual conferences of the International Society for the Scientific Study of Subjectivity. The 13th conference is scheduled to be held in October 1997 at the Newhouse School of Public Communications, Syracuse University.

References

Atkinson, J.R. (1992). QMethod (Version 1.0) [Computer software]. Kent, OH: Computer Center, Kent State University.

Brown, S.R. (1980). Political subjectivity: Applications of Q methodology in political science. New Haven, CT: Yale University Press.

Brown, S.R. (1993). A primer on Q methodology. Operant Subjectivity, 16, 91-138.

Dennis, K.E. (1986). Q methodology: Relevance and application to nursing research. Advances in Nursing Science, 8(3), 6-17.

Dennis, K.E. (1990). Patients' control and the information imperative: Clarification and confirmation. Nursing Research, 39, 162-166.

Dennis, K.E., & Goldberg, A.P. (1996). Weight control self-efficacy types and transitions affect weight-loss outcomes in obese women. Addictive Behaviors, 21, 103-116.

Larson, P.J. (1984). Important nurse caring behaviors perceived by patients with cancer. Oncology Nursing Forum, 11(6), 46-50.

McKeown, B.F., & Thomas, D.B. (1988). Q Methodology. Newbury Park, CA: Sage.

Mrtek, R.G., Tafesse, E., & Wigger, U. (In press). Q-methodology andsubjective research in pharmacy. Journal of Social and Administrative Pharmacy.

Sell, D.K., & Brown, S.R. (1984). Q methodology as a bridge between qualitative and quantitative research: Application to the analysis of attitude change in foreign study program participants. In J.L. Vacca & H.A. Johnson (Eds.), Qualitative research in education (Graduate School of Education Monograph Series) (pp. 79-87). Kent, OH: Kent State University, Bureau of Educational Research and Service.

Stainton Rogers, R. (1995). Q methodology. In J.A. Smith, R. Harre & L. Van Langenhove (Eds.), Rethinking methods in psychology (pp. 178-192). Thousand Oaks, CA: Sage.

Stainton Rogers, W. (1991). Explaining health and illness: An exploration of diversity. London: Harvester/Wheatsheaf.

Stephenson, W. (1953). The study of behavior: Q-technique and its methodology. Chicago: University of Chicago Press.

Wigger, U., & Mrtek, R.G. (1994). Use of Q-technique to examine attitudes of entering pharmacy students toward their profession. American Journal of Pharmaceutical Education, 58, 8-15.

Biographical Sketch

Steven R. Brown is Professor, Department of Political Science, Kent State University, where he teaches and conducts research in the areas of political psychology, policy, and methodology. Email address: sbrown@kent.edu.