Correspondence to: Rod Moore, DDS, MA, Department of ... their symbolic association with affect or meaning ... American may seek 'hot' liquids for a sore throat,.
Pain, 34 (1988) 195-204 Elsevier
195
PAI 01257
Basic Section Ethnographic methodologic assessment of pain perceptions by verbal description Rodney A. Moore
* and Samuel F. Dworkin
**
* Royal College of Dentistry, 8OQOAarhus C (Denmark), and * * University of Washington, Seattle, WA 98195 (U.S.A.) (Received
26 October
1987, revision received 23 February
1988, accepted
4 March
1988)
This study was designed to use known anthropologic methods to gather and analyze qualitative data about verbal descriptors of pain among 25 Chinese, and 60 Western subjects (25 Anglo-Americans and 35 Scandinavians). The sample consisted of 54 patients and 31 dentists. Key pain descriptors from each cultural context were selected for construction of pain assessment instruments which allowed multidimensional statistical techniques to translate these data into cross-cultural quantitative indices. Results revealed dimensions of pain which were universal in all cultures examined. These included time, intensity, location, quality, cause and curability. More culture-specific dimensions included the Chinese concept stuntong, a multidimensional concept of bone, muscle, joint, tooth and gingival pain. ‘Real’ and ‘imagined’ pains were contrasts described by Western subjects, especially dentists; ‘imagined pain’ being the conversion of fear or anxiety into perceived pain. These data indicate that the data gathering and data analytic methods were reliable and sensitive to cultural variables and that ethnicity played a stronger role in determining perceptions of pain description than professional socialization for this population sample of Chinese and Western subjects.
S-=-Y
Key words:
Pain description;
Pain perception;
Ethnicity;
Professional
Introduction Meaning, especially emotional significance, greatly influences the quality and quantity of pain reported, and weighs heavily in the description and measurement of pain [l-4]. There has been considerable attention paid to verbal descriptors of pain in the literature. Melzack and Torgerson [25] validated sensory, affective and evaluative classes of pain descriptors, creating the widely used McGill Pain Questionnaire (MPQ) [24]. They acknowledged that these descriptor scales, de-
Correspondence to: Rod Moore, DDS, MA, Department of Child Dental Health and Community Dentistry, The Royal Dental College, Vemrelyst Boulevard 9, DK-8000 Aarhus C, Denmark. 0304-3959/88/$03.50
0 1988 Elsevier Science Publishers
socialization
veloped in the English language in Canada, might vary according to culture, but did not specify how. However, assessment of the meaning of pain remains elusive, as recent work with Finnish, French, Argentinian, Lithuanian, Italian and Chinese versions of the MPQ have demonstrated [5,17,23,27, 29,301. A major difficulty in comparing verbal descriptors of pain among varying cultural groups seems to arise because idiomatic pain expressions are often not parsimoniously translatable from English. The intentionality of pain descriptors, not their direct translations, appear to be the key to their symbolic association with affect or meaning and this symbolic association of pain terms can be expected to vary across different social contexts. Thus, conventional social survey methods used without regard to the nature of the specific social context of subject responses may not reveal al-
B.V. (Biomedical
Division)
ternative meanings given by different cultural groups to the same stimulus situation [7,18,35]. It may be unrealistic, for example, to expect to capture valid and reliable meanings or significance assigned to a pain when the assessing inst~ment does not include response categories the patient might prefer to use a problem inherent in social survey methods using a standardized, verbatim translated verbal pain descriptor scale. The culture of the listener, e.g., physician, dentist, or pain clinician, may determine a set of meanings for similar verbal expressions that are different from the patient’s. For example, an AngloAmerican may seek ‘hot’ liquids for a sore throat, meaning thermally hot, while an Asian may seek ‘hot’ foods, e.g., ginger root in the sense of seeking to balance yin and yang forces within the body. There are identifiable and consensually validated [32] social group thought processes that have evolved historically within particular societies. It appears reasonable that in this way, the pain experience undergoes culture-specific cognitive sorting [14], resulting in types of perceptual classifications which often have different meanings, referents and intentions [15]. If these semantic sorting mechanisms indeed exert appreciable influence on the inte~retations of pain communications, it is important in researching these phenomena to be wary of methodologic assumptions and linguistic distinctions borrowed from cultures other than those of the subjects under investigation [35]. In spite of the recognition of the importance of the multidimensional and contextual nature of the pain experience and an extensive pain literature which mentions cultural variables, there is, in fact, a paucity of research on the methodologic and contextual issues relating pain experience to cultural context. In a seminal study, Zborowski [40] used participant observation methods and interviews of individu~s from 4 ethnic groups in a New York City hospital setting to focus on reactions to different kinds of postoperative pain and coping. He concluded that qualitative differences existed among this sample of Italians, Jews, Irish and ‘old’ (multi-generation) Anglo-Americans. The Irish tended to be deniers and ‘old Americans’
optimistic belittlers of pain. Italians tended to he non-optimistic profuse expressors of pain, while Jews, though equally expressive, tended to be more optimistic. Further, he reported: (1) similat observable reactions to pain demonstrated by members of different ethno-cultural groups did not necessarily reflect similar attitudes about pain or the meaning of the pain experience, and (2) similar behavioral response patterns to pain served different social interactive intentions in various cultures (e.g.. sympathy seeking vs. beliefs in ridding self of pain). Zborowski suggested that physicians use knowledge of these culturally influenced response patterns to facilitate patients’ pain coping within their own cultural context. Subsequently, empirical studies of these same ethnic groups have largely confirmed Zborowski’s findings [21,22,33,41]. The main criticism of the Zborowski work 136,381 has been the unsystematic nature of his study design in which he disregards discussions of social variables such as differences by generation, sex and social status as well as other more subtle nuances such as clinical contexts and ethnic identification. In a quasi-experimental study of dental pain and anxiety among Black, Caucasian and Puerto Rican ethnic groups, Weisenberg et al. 1371 used an 8-item questionnaire developed by Zola [41] for Italian, Irish and Anglo-American samples to measure denial of pain or willingness to deal with pain. While the Puerto Rican sample scored significantly higher on denial ratings than the other 2 groups, Weisenberg failed to observe differences between Black and Caucasian Americans. This may be due to the failure of his measu~ng instrument to validly reflect the culturally influenced pain experiences of some of his experimental groups, as argued earlier. In an attempt to capture pain-related contextual nuances across different cultural groups, the present study employed a subject centered ethnographic approach which is more similar to Zborowski’s approach than Weisenberg’s. followed by a rigorous quantitative phase to refine further comparative data collections and allow powerful statistical analyses. Our goal was to demonstrate the feasibility and validity of a new approach to the cross-cultural study of pain which
197
addresses some of the problems inherent in this area. In the study described below, we have attempted to improve on the Zborowski ethnographic interview method by matching subjects across groups by major social variables and ethnic identifications. We also used subject informants that were not experiencing pain at the time of the interviews, so that the ‘thought’ of the kinds of pains inquired about could be captured from a social group context. We started with ‘man-onthe-street’ qualitative interviews in lay language and selected key descriptors from each cultural context for instrument construction so that multidimensional statistical techniques could be used to translate the data into cross-cultural quantitative indices. The aim was to reflect the semantic validity of the qualitative data while also improving reliability and replicability. We contend along with others [16,31] that combined use of qualitative and quantitative methods may be the most appropriate means for enhancing validity and reliability of psychosocial evaluations of clinical pain behavior. The primary aim of this initial investigation was to explore the sensitivity of these methods for comparing pain descriptions across cultural contexts using ethnicity and professional socialization as variables. We used a health care communication model developed by Kleinman [20] and Chrisman and Kleinman [9] as the basis of our design. This model asserts that there are professional and popular ways of thinking about health care which can be judged to be congruent or incongruent. Dentistry was used since it is associated with many painful interactions in the patient/healer relationship, lending an opportunity to observe, in addition to exploring ethnic perceptions, how dentists and patients may agree or disagree regarding the significance of dentally related pains. Specifically, we hypothesized that the methods we used would be able to demonstrate that: (1) ethnic patient groups vary in their perceptions of pain description; (2) dental professionals differ from patients in their perceptions of pain description, regardless of ethnicity; (3) semantic measures of pain based on AngloAmerican pain perceptions are not always valid in other cultural contexts.
Methods Sample characteristics A sample of 85 subjects (54 patients and 31 dentists) were interviewed by 2 well-calibrated field researchers in the ethnic language of the interviewee. The first author was responsible for 35 Scandinavian and 25 English interviews and a Chinese assistant for 25 Mandarin interviews. There were 7 subgroups representing immigrant or first generation Anglo-American, Swedish, Danish and Mandarin Chinese subjects. Recruitment of patient subjects occurred largely through contacts with local ethnic clubs or church groups. Dentists were recruited through dental association directories in the Greater Seattle area. Each of the patient groups were demographically matched as closely as possible to reduce the number of intervening social variables other than ethnic background. Equal numbers of male and female patient subjects were chosen and matched by education (completed at least high school), age (30-60 years) and income ($S,OOO-$24,000 annually). Dentist subjects were predominantly male in this stratified sample of convenience. Study protocol (1) In open-ended interviews, subjects were asked to answer in their native language, the following questions: ‘What kinds of pain are there?,’ ‘What kinds of pains can one feel in the face and mouth?,’ ‘ What kinds of pains can one feel at the dentist?,’ and ‘What kinds of ways are there to get rid of or ease these pains?.’ The data were collected using the Metzger/Williams [26] fieldnote technique. (2) Key pain descriptors chosen from the interviews were typed onto index cards. ‘Kinds of pain’ and adjective descriptor stimuli were selected by the following criteria: (a) The stimuli are all recognizable words, at least to American non-professionals. (b) Words and terms were initially translatable to English from the 3 non-English languages and back using standard language dictionaries as reference. A system of ‘back-translations’ [6] was also used in which the words were first translated by bilingual judges and then retranslated back into the original language by judges unknown to
198
the first, thus verifying semantics. (c) Terms incorporate recognizable elements of the McGill Pain Questionnaire categories of pain description. (d) Sixty percent of the pains were also selected to be located in orofacial areas. These card instruments were numbered and presented to the same informants with instructions to sort and pile the pain descriptors by similarity, using their own criteria. (3) Reasons and meanings for sorting choices were elicited, pile by pile, in an interview after the sorting task was completed. (4) Using the key pain descriptors obtained from step 2, a row/column matrix instrument was constructed and administered for ‘kinds of pain’ (nouns) vs. adjective ‘descriptors.’ Subjects were
TABLE
I
ITEMS CHOSEN
FOR INSTRUMENT
CONSTRUCTION
Pains English
Descriptors
Joint pain Common headache Migraine Sinus headache Backache Toothache Child birth labor pain Muscle pain Mouth blister pain Loss of loved one Burn pain Electric shock pain Pain of holding jaw open Broken arm or leg pain Pain of air on open tooth cavity Hit finger pain Pain after operation Needle injection (shot) pain Stomache ache Tooth drilling pain Pain bruise Kidney stone pain Pain after tooth extraction Tooth cleaning pain Heart pain Chest pain Sourish pain Menstrual pain Sore throat pain Gum pain Deep bone pain
Stinging Tingling Burning Pulling Cramping Gripping Pinching Crushing Tearing Sharp Stabbing Continuous Intermittent Throbbing Sympathy producing Dull, aching Sore Quick, short lasting Tender Swelling Terrible Intense Unpleasant Excruciating Irritating Good pain Numbing Sour&h
English
instructed to make an entry cells with a number where plied to a kind of pain. The English version of instrument items used in Table I.
into applicable matrix a pain descriptor apthe card the study
and matrix is listed in
Data analytic methods (a) Open-ended data resulting from responses to the ‘what kinds of pain are there?’ questions were content analyzed by frequency tabulations of words and/or phrases. Many subjects volunteered supplemental statements that were helpful in assigning kinds of pain and remedies to categories or domains. (b) Pile sort data responses were organized into taxonomic ‘trees’ for each group. The trees reflect perceived similarity among kinds of pain. These structural comparisons of pain were generated using an algorithm developed by Burton and Romney [8] and a hierarchical clustering computer program (ALPAIR) devised by D’Andrade [12]. The ALPAIR program uses the Mann-Whitney U statistical test, a non-parametric measure of association to determine non-chance allocations among pain descriptors. ALPAIR judges the cognitive proximity of all possible pairs of cards sorted by subjects. From the history of repeated comparisons in the clustering process the taxonomic tree structure is constructed using the ALPAIR statistical criteria. This represents the strength of associations among words from the enormous number of comparisons (see Figs. 1 and 2 for examples of pile sort ‘trees’). The cluster nodes of the tree structure indicate levels of strengths of associations between items. Associations decrease, or become weaker, as the nodes ascend the ordinate axis. Reasons given for pile sorting choices, elicited separately from all subjects, were content analyzed by word and phrase frequency to further clarify and explain tree structures. (c) Pain descriptor matrix data were used to cross-validate open interview and pile sorting results and to generate multidimensional cluster configurations for group comparisons. These data allow the statistical reliability of this ethnographic methodological approach to be evaluated, as well
199
ANGLO-AMERICAN
DENTISTS (ADI
-t II
I
I
-
1
b
I
”
I”
Ill
Vll,
VII
“I
Fig. 1. Tree structures generated from pile sorting clusters of American dentists.
CHINESE PATIENTS &PI I” = 151
I
I__‘_8
I
II
III
IV
V
VI
“II
Fig. 2. Tree structures generated from pile sorting clusters of Chinese patients.
VIII
XI
200
as allowing comparison of individual responses with those of the whole population. Multidimensional scaling (MDS) procedures [39] were also employed to analyze the matrix data. First, a measure of profile similarity was obtained between each row of all of the subjects’ matrices using product-moment correlation coefficients. This generated new matrices of monotonically transformed similarity values suitable as input for a computer program that positioned all 85 subject correlates onto a 2-dimensional plane by proximity of pain description (see Fig. 3 below), The multidimensional program used here was ALSCAL [34]. ALSCAL computed weighted values from the correlation matrices using an alternating least squares parameter (ALS) estimation method [13] and generated a series of successive iterations. Convergence of values and coordinates on a Euclidean plane were calculated and printed in relation to the overall similarity of the 85 data
points for the set of matrix choices. The coordinate values and axes of the planes in Fig. 3 serve merely as spatial orientation for determining similarity relationships between points and have no other meaning or purpose. The numerical values refer to membership in the 7 subgroup populations. In the assessment of reliability [lO,llj, meaningful scaling requirements are met when 3 conditions are satisfied: (1) a fairly consistent set of data relations within groups are obtained (i.e., congruity); (2) that a mathematical model can be selected, which contains variables that can be matched to the empirical conditions, and a solution procedure exists for fitting MDS (ALSCAL) model (i.e., fit); and (3) demonstration of a meaningful relationship of the obtained scale to other data or scales that are external to the measurement procedure (i.e., generalizability).
Results and discussion 2.5
t
SCANDINAVIAN PATIENTS ,
2.5
2.0
2.0
1.5
1.5
1.0
1.0
0.5
0.5
0
0
~0.5
-0.5 AND DENTISTS
-1.0
-1.0 15
~1.5
DENTISTS
-2.0
-2.0
2.5
-2.5
1
2.5
~2.0
-1.5
1.0
PAIN
0.5
0’
0.5
1.0
1.5
2.0
2.5
DESCRIPTORS
Fig. 3. Two-dimensional representation of proximity of individuals by their perceptions of pain descriptors derived from ALSCAL procedures. Legend codes: 1 = AD = AngloAmerican dentists; 2 = AP = Anglo-American patients; 3 = CP = Chinese patients; 4 = CD = Chinese dentists; 5 = SP = Swedish patients; 6 = ScD = Scandinavian dentists; 7 = DP = Danish patients.
Open-ended interviews Differences in pain descriptions as among ethnic patient groups and between patients and dentists were not great in quantity, but the frequencies with which certain pain concepts such as suantong (a unique Chinese descriptor) were reported clearly showed qualitative differences, especially between Western and Chinese groups. Differences between patient and dentist perceptions of kinds of pain and descriptors were not as great as expected. A related study [28] indicates that patients and dentists differed in descriptions of pain coping remedy to a much greater extent than they did for verbal descriptions of pain. Pile sorts Figs. 1 and 2 illustrate ‘tree’ results for AngloAmerican dentists (AD) and Chinese patients (CP), respectively. Clustering results for the AD subgroup (n = 10) show the 32 ‘kinds of pain’ stimuli to fall into 8 clusters. Cluster 1, the largest, is characterized by a muscular/bone locus, cluster II by pains never experienced, cluster III pains which vary in time and intensity, cluster IV by burning pains of oral soft tissues, cluster V by pains of
high intensity and cluster VII by external source (e.g., dentist) provoked. VI (loss of loved one) and VIII the unknown ‘type Q pain,’ each contain a single unassociated pain. Clustering results for the CP subgroup (n = 15) show that the same 32 pain terms fall into 10 clusters. Cluster I contains pains with a muscular/bone locus in addition to the not previously described concept called suantong (‘sour&h pain’). Cluster II is characterized by head pains, cluster III by tooth and gum pains, cluster IV by pains of the abdomen, cluster V by ‘painful when touched or moved,’ cluster VI by intensity and time factors, cluster VII by pains related to the chest (including mental anguish), cluster VIII by injurious pains of bones, and IX by a nonsense pain (type Q pain). AD and CP subgroups typify East-West differences, but all subgroups universally clustered on the dimensions of muscle/bone locus, chest and/or abdominal locus, mouth-tooth locus, head locus, intensity, time, quality, cause and curability. Only Chinese subjects classified pains according to the suuntong dimension, whereas Western subjects related ‘sourish pain’ to concepts of oral and gastrointestinal burning pain. Swedish and Chinese patients linked ‘loss of loved one’ with heart and chest pain consistently. All other groups sorted it as a dimension of its own. ‘Loss of loved one’ was included in an attempt to explore ‘emotional pains.’ Nearly all groups said they sorted the pain stimuli according to location, time, intensity, quality, curability and cause dimensions. In addition to these, Western subjects named a mental/physical dimension, differentiating emotional (‘imagined’) or physical (‘real’) pains. The concept of ‘real vs. imagined’ pain (mental/physical dichotomy) during routine dental procedures was more often named by dentists than patients. This mental/physical dichotomy was barely mentioned among the Chinese subjects. The unique ‘suantong’ pain appeared to be most similar to Western ‘quality’ and ‘cause’ dimensions and was specific to certain body areas and depths. These Chinese experienced suantong in tooth drilling, tooth cleaning and in muscle, bone and joint pains. Tooth drilling suantong, for example, was described as being a dull, short
lasting pain, in contrast with Westerners who described dental pain to be sharp and intense. One Chinese subject described suantong as ‘pain in the bones, both itchy and painful - like there are hundreds of needles stabbing.’ Matrix data Aggregate values set in contingency table form for each subgroup indicated corroboration with open-ended and pile sort data. Proximity or distance between the points and relationships of clusters of points in the MDS configuration (Fig. 3) indicates that Western dentist groups and Western patients revealed similar cognitive patterns for the description of pain. The Chinese dentists, however, indicated closer ethnic affiliation with Chinese patients in their perceptions of pain description, than they did with Westem dentists. Discrete ethnic patient groups were identified. Certain examples make these differences apparent, such as the observation that ‘gripping’ pain has an immediate association with a physical grip in Anglo-American context, while in Scandinavian contexts the immediate association is that of an emotionally moving experience. Some outliers from each group approximated other groups. Reliability Reliability measures of congruity for these matrix data were obtained from the normalized input correlation matrices prior to ALSCAL treatment. Ethnic groups had high intra-group reliability (Cronbach’s alpha = 0.970 for descriptors) on the average for each group (P < 0.05). The measures of fit are so-called ‘stress’ levels of how the data fit the ALSCAL program without damaging the interpoint relationships on the Euclidean projection. These ‘stress level’ measures and squared correlation coefficients (RSQ) were computed by the ALSCAL program (via SPSS-X, version 2.2 and SAS). RSQ values are the proportion of variance of the scaled data (which measures disparities in the correlation matrices) which is accounted for by the distances between points. Results in this case show an excellent fit between the data and the solution since Kruskal’s SSTRESS formula 1 = 0.204 (low stress) and RSQ
202
= 0.898 (high reliability) for the descriptors. Thus, the empirical data used to determine the relationships among pain terms fits the theoretical statistical model extremely well. These findings, therefore, lend support to the results of the qualitative data gathered in openended interview analyses and support the face validity of ‘reasons given for sorting choices’ of the pile sort data. Agreement with these data (generalizability) plus finite group identifications by individual proximity on the Euclidean configurations (further evidence of congruity) indicate that this methodologic approach appears to be reliable as well as internally valid.
Conclusions The hypotheses regarding the influence of cultural variables on pain description were largely affirmed, with only one exception. Dental professionals and patients of the same subgroup were not confirmed to be different in their perceptions of pain description in any of these data. Ethnicity therefore seemed to be a more potent cultural variable than professionalization for understanding meanings given to pain descriptors for ?his sample. The results suggest that the relative influence of ethnic membership and professional/ lay membership depends on the cultures compared. since there were greater East-West differences than Anglo-American/Scandinavian differences. Scandinavian and Anglo-American groups showed a strong influence of professionalism on pain descriptor meanings; both dentist groups were more similar to each other than to their respective patient populations which, in turn, differed from each other. A comparison of these Western groups with the Chinese groups, however, showed a greater influence of ethnicity. The study also demonstrated that the ethnographic methods used appeared to be sensitive to differences in the ways people (dentists and patients) from different ethnic subgroups ascribe meaning to pain descriptors. Some components of the sociocuhural descriptive data, such as pile sort and matrix data lent themselves more easily to quantification and statistical analysis. The main
strength of the pile sort methods was their ability to reveal the tacit cultural organization of pain description. Matrix data methods provided quantitative cross-validations of the other methods, assessed cognitive proximity of individual responses and verified consistency among culturai groupings, thus enhancing reliability of the methods. Other methods, such as open-ended interviews, taken alone. may have questionable reliability, but their richness and concreteness in description lend semantic credibility and face validity to subsequent analysis with more quantitative data gathering instruments. These early findings may not be generalizable to entire cultural groups. Results with this small sample do indicate, however, that many of the same words used by subjects in different ethnic groups denote different symbolic meanings, and that ethnicity influences perception of pain, e.g.. ‘gripping’ pain in American and Scandinavian contexts. To some degree, ethnic differences we observed may also reflect the pervasiveness of cultural variables across common demographic characteristics, e.g., professional and lay groups. Further research with the methods used here are needed, but it appears from the results of this study that researchers and clinicians must design or interpret data from questionnaires such as the MPQ with caution. Any test instrument of this kind, developed and validated in one culture. should not necessarily be expected to be valid in another context. Even if verbal pain descriptors yield similar scale scores, these scores may denote different perceptual experiences of pain, as noted earlier with gripping pain. Therefore, it has become necessary to regenerate and/or improve the process which was used to develop and validate the MPQ to find the semantic equivalents in other languages, rather than using literal translations of the English version. The advance that the MPQ represents is the introduction and validation of the multidimensional scaling of pain experience. No single dimension, such as stimulus or affective intensity of pain experience, is adequate to capture specific cultural meanings of pain. Multiple combinations of various dimensions of pain are important to the meaning of the pain in the patient’s contextual
203
reality, such as the existence of the suuntong concept in a Mandarin Chinese context and the meaning, for example, it lends to tooth drilling pain. This supports a clinical perspective in which the clinician needs to understand underlying assumptions made in the pain description process. Clinicians must be aware of their own limitations and biases when evaluating patients in pain who have other cultural backgrounds than their own. The results of this study suggest that ethnographic methods may provide much needed coverage of contextual differences in the research of pain descriptions and perceptions.
Acknowledgements Supported by Grants T32-DE-07132 and RR05346 from National Institutes of Dental Research, Bethesda, MD, U.S.A.
References
5
6 7
8 9
10
Beecher, H.K., Relationship of significance of wound to pain experienced, J. Am. Med. Ass., 161 (1956) 1609-1613. Beecher, H.K., The measurement of pain, Pharmacol. Rev., 9 (1959) 59-209. B-her, H.K., Measurement of Subjective Responses, Oxford University Press, New York, 1959. Bonica, J.J., Management of Pain, Lea and Febiger, Philadelphia, PA, 1953. Boureau, F., Luu, M. and Doubrere, J.F., Qualitative and quantitative study of French pain McGill adapted questionnaire in experimental and clinical conditions, Pain, Suppl. 2 (1984) S422. Brislin, R.W., Lonner, W.J. and Thomdike, R.M., CrossCultural Research Methods, Wiley, New York, 1973. Brown, E.D. and Se&rest, L., Experiments in cross-cultural research. In: H.C. Triandis and J.W. Berry (Eds.), Handbook of Cross-Cultural Psychology. Vol. 2: Methodology, Allyn and Bacon, Boston, MA, 1980, pp. 297-318. Burton, M.L. and Romney, A.K., A multidimensional representation of role terms, Am. Ethnol., 2 (1975) 397-407. Chrisman, N.J. and Kleinman, A., Popular health care, social networks and cultural meanings: the orientation of medical anthropology. In: D. Mechanic (Ed.), Handbook of Health, Health Care and Health Professionals, The Free Press, New York, 1983, Ch. 26. Cliff, N., The idealized individual interpretation of individual differences in multidimensional scaling, Psychometrika, 33 (1968) 225-232.
11 Cliff, N., Scaling, Ann. Rev. Psychol., 24 (1973) 473-506. R.G., U-statistic hierarchical clustering, Psy12 D’Andrade, chometrika, 43 (1978) 59-67. 13 De Leeuw, J., Young, F. and Takane, Y., Additive structure in qualitative data: an alternating least squares method with optimal scaling features, Psychometrika, 42 (1976) 471-503. pain semantics, Pain, 9 (1980) 14 Diller, A., Cross-cultural 9-26. 15 Fabrega, Jr., H. and Tyma, S., Culture, language and the shaping of illness: an illustration based on pain, J. Psychosom. Res., 20 (1976) 323-347. 16 Glaser, B. and Strauss, A.L., The Discovery of Grounded Theory, Aldine, Chicago, IL, 1967. 17 Hui, Y.L. and Chen, A.C.N., Headache in Taiwan: assessment and characterization by a Chinese version of McGill pain questionnaire and headache topography chart, Pain, Suppl. 2 (1984) S590. 18 Kirk, J. and Miller, M.L., Reliability and validity in qualitative research. In: J. Van Maanen (Ed.), Qualitative Research Methods, Ser. No. 1, Sage, Beverly Hills, CA, 1986. 20 Kleinman, A., Patients and Healers in the Context of Culture. An Exploration of the Borderland between Anthropology, Medicine and Psychiatry, University of California Press, Berkeley, 1980. 21 Koopman, C., Eisenthal, S. and Stoeckle, J.D., Ethnicity in the reported pain, emotional distress and requests of medical outpatients, Sot. Sci. Med., 18 (1984) 487-490. J.J., Ethnicity and the pain 22 Lipton, J.A. and Marbach, experience, Sot. Sci. Med., 19 (1984) 1279-1298. 23 Maiani, G. and Sanavio, E., Semantics of pain in Italy: the Italian version of the McGill Pain Questionnaire, Pain, 22 (1985) 399-405. major prop24 Melzack, R., The McGill Pain Questionnaire: erties and scoring methods, Pain, 1 (1975) 277-299. 25 Melzack, R. and Torgerson, W.S., On the language of pain, Anesthesiology, 34 (1971) 50-59. 26 Metzger, D. and Williams, G., Tenejapa medicine. I. The curer, Southeast J. Anthropol., 19 (1963) 216-234. 27 Molina, F.J., Molina, B.Z., Molina, J.C. and Coppo, C., The Argentine pain questionnaire Pain, Suppl. 2 (1984) S42. 28 Moore, R.A., Miller, M., Weinstein, P., Dworkin, S.F. and Liou, H.-h., Cultural perceptions of pain and pain coping among patients and dentists, Commun. Dent. Oral Epidemiol., 14 (1986) 327-333. validation of the Lithuanian 29 Pakula, A., Psychophysical pain questionnaire (LPQ) sensory scale: a clinical model, Pain, Suppl. 2 (1984) S43. H., Verbal measurement in 30 Pontinen, P.H. and Ketovuori, non-English language: the Finnish pain questionnaire. In: R. Melzack (Ed.), Pain Measurement and Assessment, Raven Press, New York, 1983, pp. 85-93. 31 Reichardt, C.S. and Cook, T.D., Beyond qualitative versus quantitative methods in evaluation research. In: T.D. Cook and C.S. Reichardt (Eds.), Research Progress Series in Evaluation No. 1: Qualitative and Quantitative Methods in
204
32
33
34
35
Evaluation Research, Sage, Beverly Hills, CA, 1979, pp. 7-32. Romney, A.K. and Weller, B., Culture as consensus: a theory of culture and informant accuracy, Am. Anthropol., 88 (1986) 313-338. Sternbach, R.A. and Tursky, B., Ethnic differences among housewives in psychophysiology and skin potential responses to electric shock, Psychophysiology, 7 (1965) 241-246. Takane, Y., Young, F. and De Leeuw, J., Nonmetric individual differences in multidimensional scaling: an altemating least squares method with optimal scaling features, Psychometrika, 42 (1977) 7-67. Trimble, J.E., Lonner, W.J. and Boucher, J.D., Stalking the wily emit: alternatives to cross-cultural measurement. In: S.H. Irvine and J.W. Berry (Eds.), Human Assessment and Cultural Factors, Plenum, Boston, MA. 1983, pp. 259-273.
36 Weisenberg, M., Cultural and ethnic factors in pam. In: 1. Al-Issa (Ed.), Culture and Psychophathology, University Park Press, Baltimore, MD, 1982, pp. 415-432. 37 Weisenberg, M., Kreindler, M.L., Schachat, R. and Werboff. J., Pain: anxiety and attitudes in black, white and Puerto Rican patients, Psychosom. Med., 37 (1975) 1233135. 38 Wolff, B.B. and Langley, S., Cultural factors and the response to pain: a review, Am. Anthropol.. 70 (1968) 494-501. 39 Young, F.W. and Hamer, R.M. (Eds.), Multidimensional Scaling History, Theory. and Applications. Lawrence Erlbaum, Hillsdale, NJ, 1986. 40 Zborowski, M., People in Pain, Jossey-Bass, San Francisco, CA, 1969. 41 Zola. J.K., Culture and symptoms: an analysis of patients presenting complaints, Am. Sot. Rev., 31 (1966) 615-630.
