Determinant service attributes in the formulation of attitudes toward rehabilitation facilities.

All developed countries have a large service sector. In the United States the provision of services has displaced the production of goods as the principal economic activity (Riddle, 1986). Rehabilitation, as one part of the healthcare service industry, has become an important sector for healthcare service providers (Walker, 1988). In the early 1980s, rehabilitation services were largely marketed to organizations that were legally or ethically responsible for the rehabilitation of people with disabilities (Nadolsky, 1984). Since then the nature of the marketing activities of rehabilitation providers has changed with the growth of a private rehabilitation sector as a result of the budget policies of the early 1980s that reduced financial support for federal and state rehabilitation programs, and the passage of legislation that mandated the public rehabilitation program to serve individuals who are more severely disabled (Hasbrook, 1981; Matkin, 1980).

As new private service providers continue to enter into the rehabilitation market, especially those growth sectors in pediatrics, geriatrics and occupational/industrial area (Walker, 1988), the adoption of market concepts and strategies gives the service provider an edge over competitors. Driven by the competitive force, the nature of healthcare service delivery has become more consumer-oriented (Berkowitz & Flexner, 1980-81). It is widely recognized that consumers are now playing a more significant role in choosing their own healthcare services (Berkowitz & Flexner, 1980-81; Boscarino & Steiber, 1982; Smith & Clark, 1990). There are also calls for increased accountability and responsiveness of service providers (Habeck, 1993; Patterson & Leach, 1987; Patterson & Marks, 1992). One measure used to monitor accountability is the evaluation of consumer satisfaction. This would seem to be the natural outgrowth of both a consumer-oriented society and a national interest in evaluation and accountability (Patterson & Leach, 1987).

Another accountability measure used is the evaluation of consumers' attitudes. Despite the fact that a causal relationship between the consumer's attitudes toward a service and his/her purchase outcome has not been established, it is generally agreed that attitudes exert both a directive and a dynamic influence on behavior (Block & Roering, 1979). Because attitudes influence behaviors, the concept of attitude has been one of the most prevalent and important concepts in the study of consumer behavior (Wilkie & Pessemier, 1973).

There are two approaches to the understanding and measurement of attitudes (Hooley, 1978). One approach is to determine the overall attitude and then identify the reasons for or the components of this overall attitude. Another approach is to examine the individual components or the attributes and aggregate these findings to form the overall attitude using summative attitude models (Ajzen & Fishbein, 1980; Bass & Talarzyk, 1972). Healthcare academicians and marketeers have adopted these two approaches to study healthcare consumers' attitudes towards hospitals (e.g., Boscarino & Steiber, 1982; Elbeik, 1986; Smith & Clark, 1990; Woodside, Nielsen, Walters & Muller, 1988) and other types of medical services, such as outpatient clinics (Lim & Zallocco, 1988), ambulatory healthcare centers (e.g., Klegon, Kingstrom & Gregory, 1982), nursing homes (Lim & Zallocco, 1988), and specialty clinics (Flexner, McLaughlin & Littlefield, 1977).

Elbeik (1987) identified a total of 103 service attributes used by consumers in selecting their ideal hospitals. Their importance ranged from critical to very unimportant. Smith and Clark (1990) provided a comprehensive review of empirical findings of 10 determinants of hospital and medical facilities image, that including quality of physicians, treatment equipment, diagnostic equipment, overall care quality, interpersonal care, staff awareness of patient's personal needs, patient control of hospital experience, patient information and education, cost of hospital stay, and convenience of location.

Unfortunately, literature specific to rehabilitation facilities is scarce. The current study on which this paper is based adopted an aggregate approach (Hooley, 1978) in studying rehabilitation consumers' attitudes. This approach focuses on the identification of rehabilitation consumers' service attributes that are important in explaining preference (affective element of attitude) and in predicting behaviour (the conative element of attitude) of the consumer.

The overall purpose of the study was to gain a fuller understanding of the rehabilitation service market in terms of determinant attributes used by consumers in perceiving and selecting among different rehabilitation services. The study, which was conducted in Hung Kong, consisted of three parts: a) the development of a set of service attributes and attribute factors; b) the identification of different segments of consumers based on their socio-demographic variables; and c) the identification of consumers' perceptions and preferences of the services provided by three groups of rehabilitation providers. This paper presents the findings of the first part of this study and considers implications for rehabilitation service providers.

Method

Identifying service attributes

A set of service attributes relevant to rehabilitation facilities was derived from a review of both secondary and primary data. Because information on service attributes used by consumers in forming attitudes towards rehabilitation facilities was scarce, literature on a range of medical facilities was reviewed. Over a hundred service attributes were identified in some studies (e.g., Elbeik, 1987), it was important to determine those attributes relevant to rehabilitation facilities. Sharpe and Granzin (1974) identify several directions for defining relevance. These include looking at each attribute in terms of its importance in the formulation of overall perceptions; the identification of differences between services; in determining preferences for services; and in determining behavior. This study adopted the last two directions.

Primary data were derived using two unstructured methods (Hughes, 1974): a focus group and a set of individual interviews. Focus group is a widely used method in which a small number of individuals are brought together and encouraged to talk about a subject rather than to answer yes or no to specific questions. In terms of stimulating new ideas (Goldman, 1969), provoking greater spontaneity (Chisnall, 1981), and identifying the strength of an attitude (Goldman, 1969), focus group is a superior method than individual interview. In general, the interview method, either individual or group, has three distinctive advantages: it is the most flexible data collection method; it involves participants who can address the issue of concern; and the response rate is high in comparison with other methods (Boyd, Westfall & Stasch, 1985; Day, 1974).

Subjects

A purposeful sample was used for this part of the study. This group was made up of consumers who had been off work for more than three days due to work injury or occupational disease and who had used the rehabilitation services provided by a local workers' health center(1) in the past six months. Because there were only twenty consumers who met the criteria, all were invited to participate in the study. Out of those twenty prospective participants, six were not contactable and two declined to participate. Of the remaining twelve, six participated in a focus group discussion and six in individual telephone interviews.

The six participants in the focus group consisted of three males and three females with ages ranging from 20 to 46; five were married; and all had sustained physical injuries, including injuries to the head, upper limbs, or lower back. Their occupations consisted of a labourer, a typist, a factory worker, a technician and two school teachers. The six participants in the individual interviews consisted of two males and four females. They were a semi-skilled worker, a skilled technician, a dancer, a butcher and two labourers; two were married; and their ages ranged from 23 to 60. All had sustained upper or lower limb injuries.

Procedure

In conducting the focus group, the investigator first explained to the participants the purpose of the discussion session, the duration of the session, and the respective role of the participants and the investigator. The introduction session was followed [TABULAR DATA FOR TABLE 1 OMITTED] immediately by a preparatory activity in which the participants were asked to identify their preferred rehabilitation facility. Participants were then asked: "What attributes would you consider when formulating your preference for the particular rehabilitation facility you have just mentioned?" The discussion ran for about two hours and the entire session was tape-recorded to allow further analyses.

A number of techniques were used to elicit responses. These included deprivation questions, "deception", sophisticated naivete, and non-directive comments (Goldman, 1969). Deprivation questions inquired into the relative value of various services. A sample question was: "Which of the following services provided by the facility would you miss most if it were no longer available to you?" Deception was used to test the participants' convictions by throwing in incorrect statements such as: "It makes no difference to your recovery as to which rehabilitation facility you attended." Sophisticated naivete involved getting the participants to explain their seemingly obvious thoughts to the investigator. Non-directive comments, such as "You seem satisfied with the service," encouraged participants to reflect further on feelings towards the services.

Lowe's "qualitative coding matrix" technique (1990) was adopted not only to analyse the data but to increase the reliability and the generalization power of the findings from the discussion. It provides a three stage framework of "open", "axial" and "selective" codes. "Open" codes are the lowest level of descriptive labels refer to raw data provided by the participants and may include a preliminary analysis by the investigator during the discussion. The "axial codes" are created by the investigator using his judgement based on the "open codes" which have logical relationships with one another. A single "axial code" is often created from a number of "open codes" which have logically similar characteristics. The highest level is "selective coding". These codes are up-graded "axial codes" that have been drawn together from other information sources. This process is aimed at producing a set of rehabilitation service attributes perceived by the participants as relevant and important.

A slightly different initial information eliciting technique was used in the six individual telephone interviews. The respondent was asked to reflect on and name five attributes that s/he would use to determine the preference for a particular rehabilitation facility. The attributes were recorded in order of recall.

A total of 32 attributes were derived and formed a checklist. A pilot study with ten respondents was conducted to test the practicability and communicability of the items. Subsequently, 6 labels ("trust in doctor", "trust in therapist", "competence of the rehabilitation staff", "rehabilitation staff explain procedures that concern you", "rehabilitation staff treat you as an individual", and "rehabilitation staff are attentive to your needs") were eliminated from the original list. The remaining 26 attribute items were used for the next stage of the study.

Determining important attributes

In the formulation of perceptions and selection of a rehabilitation facility, some attributes may be more important than others. Several approaches are available for estimating attributes importance (Alpert, 1971) and these may be classified broadly as a) direct questioning, b) indirect questioning, and c) observation and experimentation. This study adopted a direct approach which is more efficient than indirect methods in identifying important attributes for small samples (Alpert, 1971). This approach is simple to use and the cost involved is lower. Under the direct approach, the most frequently used methods in data collection are mail questionnaire, telephone interview, and personal or face-to-face interview. The advantages and disadvantages of each method are discussed in full by Erdos, Payne, and Mayer (1974) in the "Handbook of Marketing Research." Survey questionnaire by direct questioning was adopted for this study.

Subjects

For the conduction of the survey questionnaire, subjects were selected on the basis of his/her having had previous experience in using a rehabilitation facility during the past six months. The criterion ensures that all subjects were reasonably familiar with [TABULAR DATA FOR TABLE 2 OMITTED] at least one rehabilitation facility before rating the attribute items. The sample consisted of 51 injured workers attending disability assessments at two hospitals, one on the Hong Kong Island and the other on the Kowloon Peninsula. The workers, while waiting to be assessed, were randomly approached and invited to participate in the survey. All had received rehabilitation services during the past six months. The participants varied widely in age, income, education, occupation, and the type of injury sustained. The sample was heterogeneous and closely resembles the general population in need of occupational rehabilitation in Hong Kong. The subjects were predominantly male (91.8%), low income earners, and none earned more than HK$25,000 (US$3,200) per month. While most (46.9%) had received only a primary school education, others received either secondary (32.7%) or high school (20.4%) education. None of the subjects had received a college education.

Procedure

A survey questionnaire consisting of three sections was developed. The first section was composed of questions relating to socio-demographic and injury data including age, sex, marital status, family income, occupation, education level, home ownership, duration and kind of injury, resumption of duties, and duration of sick leave. The second section consisted of questions related to the types of rehabilitation facilities commonly visited by the respondents. Rehabilitation facility was defined operationally as a premises whereby treatments of work injury or occupational disease were delivered by at least one rehabilitation practitioner beside a medical practitioner. The third section contained a modified checklist of the 26 attribute items derived from the first stage of the study. A Likert summated rating was used to measure respondents' importance ratings along the attributes identified. It required the respondents to rate each item by indicating whether they considered it being a) very important, b) important, c) don't know, d) unimportant, and e) very unimportant in determining their choices of rehabilitation facilities.

In a pilot study, the questionnaire was administered to ten injured workers attending disability assessment in a regional Hospital on the Hong Kong Island to test the practicability and communicability of the questionnaire. Follow-up discussions were held to review the questionnaire. Two points were concluded from the pilot study. First, the respondents had difficulty in understanding the concept of attitude in the attitude question "How important are each of the following factors in determining your attitudes toward the rehabilitation facilities visited?" Second, the respondents encountered difficulties in differentiating some of the attitude items, especially those related to the professional and rehabilitation staff. Subsequently, three aspects of the questionnaire were modified. First, the doubtful question was re-phrased as "How important are each of the following factors in determining your choice of rehabilitation facilities?" Second, the items in the attribute checklist were written as statements.

Identifying underlying factors

The next stage, using data collected from the third section of the survey questionnaire, involved the reduction of the interrelated attributes to a smaller number of factors. As the reliability of the reduction process is sensitive to the sample size, and it was suggested that five sample cases for each observed attribute would ensure reliability (Tabachnich & Fidell, 1989), an additional sample was used, which consisted of 91 undergraduate occupational therapy students at the Hong Kong Polytechnic. Data were reduced by the Principal Components Analysis (PCA) technique, a reduction procedure belonging to multivariate statistical techniques (Dillon & Goldstein, 1984). The technique essentially transforms the original variables into a smaller set of linear combinations that account for most of the variance of the original set. The basic assumption of this technique is that underlying factors can be used to explain complex phenomena.

To summarize, this study adopted an aggregate approach in studying rehabilitation service consumer attitudes. A set of service attributes was derived from secondary data as well as from individual interviews and a focus group. A questionnaire survey was then used to determine the relative importance of these service attributes, and the PCA technique was applied to identify the underlying attribute factors.

Results

Relevant attributes

Data were processed by using Lowe's "qualitative coding matrix" technique (1990) as mentioned above. The first five service attributes mentioned (open code) by each of the six respondents participating in the individual interview are recorded in Table 1. Most items were related to the cognitive element of attitudes with some related to the affective element of attitudes. "Treatment effectiveness" and "improvement in sense of well-being" were the first and most frequently mentioned items, followed by "thorough examination" and "concern" and "care". Items with at least two responses included "equipment", "provide information", "follow-up" and "trust". Items mentioned only once included "comprehensive service", "effective management", "provide support", "waiting time", and "cost".

Attitude statements (open code) such as "I go back to my therapist even though I know that he can not cure my condition...because I know he is trying the best to help" and "people there have the heart for patients," relating to perceptions and preferences of a rehabilitation provider were recorded from the focus group discussion.

Closely related items/statements such as "concern", and those just mentioned were collated to form the axial code "caring and concern". Combining with relevant secondary data, a selective code was derived and in this example, the "caring attitude". A total of 26 attributes (selective codes) were derived from the literature, focus group and the interviews, and written into the survey questionnaire as a list of attribute items (Table 2), later converted into a set of statements.

Underlying factors

There were a total of 142 respondents (91 students and 51 injured workers) to the third section of the survey. Two questionnaires were eliminated from analysis because the respondents rated the attributes indiscriminately. Because the function of PCA is to obtain common dimensions, the ratings of each attribute must be correlated to each other for the factor model to be appropriate. The correlation matrix for the attributes was calculated [TABULAR DATA FOR TABLE 3 OMITTED] by using the SPSS/PC+ V.20 package. The results showed that ratings of all attributes except "location of rehabilitation facility (1)" and "availability of all rehabilitation services needed (4)" have correlation coefficients greater than 0.3 with at least one attribute rating in the set. The results supported the use of the factor model.

Furthermore, as the reduction process involved the use of two samples that differed in characteristics, informal procedures (Tabachnick & Fidell, 1989) were used to inspect the pattern and magnitude of the correlations between variables and factors of the two samples and the results showed that the number of factors generated were the same for the two samples, and the patterns and magnitudes of the correlations were similar.

Following the correlation procedure, the PCA technique was applied. Because there could be as many principal components as there were variables, Kaiser's approach (1958) was adopted to select those significant principal components and the criterion was to retain those components whose eigenvalues were greater than one. A clustering procedure suggested by Dillon and Goldstein (1984, p. 69) was adopted and all 26 attributes were clustered into eight factors. All eight factors had eigenvalues greater than 1 and the total variance explained by these eight factors was 66.3%. The assessment of significance was based on the statistical significance. It could be seen that all variables grouped in a cluster have statistical significant loadings greater than .36 with most of the loadings greater than .50. The factor extraction results and the labels of the eight clusters are presented in Table 3. It should be noted that the attribute label "doctors examine [TABULAR DATA FOR TABLE 4 OMITTED] you thoroughly" has been transferred from factor 7 "waiting time" with a significant loading of 0.5274 to factor 6 with a significant loading of 0.3620 as it appears obvious that the two attributes "waiting time" and "doctors examine you thoroughly" did not share a common underlying factor.

Relative importance of the attributes

The next stage involved the determination of relative importance of the 26 attributes and the 8 factors by comparisons of mean importance ratings using the Paired T-Test. Data collected from only the workers' sample were used. Table 4 shows the mean importance ratings of all 26 attributes. The most important attribute, with an importance rating of 1.53, was the "competence of therapists" and the least important attribute, with an importance rating of 3.04, is the "ability to see therapist of choice", Paired T-tests were applied to test the differences among the mean importance rating of each attribute. The mean importance rating of each attribute was significantly different from at least 7 other mean importance ratings at P [less than] 0.01 level. The mean importance rating of the label "ability to see therapist of choice" showed statistical differences from all others at the P [less than] 0.01 level.

Relative importance of the factors

Mean importance ratings of the factors (Table 4) were computed by averaging the mean importance ratings of those attributes being included in each of the factors. By comparison of the mean importance ratings, the most important factor with a mean importance rating of 1.71 was the label "DOCTORS EXAMINE AND TREAT YOU EFFECTIVELY", and the least important one with a mean important rating of 2.80 is the label "ABILITY TO SEE MEDICAL STAFF OF CHOICE".

Paired T-Tests were used to compare the mean scores of the mean importance ratings of the factors. The results are presented in Table 5. The mean score of the most important factor (6) "DOCTORS EXAMINE AND TREAT YOU EFFECTIVELY" differs significantly from the other factors except factors 4, 1 and 5. The mean scores of factors 3 and 8 differ significantly from all other factors.

Discussion/conclusion

From both the individual telephone interviews and the group discussion, the most frequently named attribute was related to the effectiveness of treatment, an increase in the sense of well-being or a lessening of the pain experienced following treatment. The author believes that the core benefit pursued by the rehabilitation consumer is a recovery to a premorbid healthy state so that the person can be independent and productive again. Consumers search for good physicians who can cure. When the market fails to deliver this core benefit, the consumer will seek out alternative ways to fulfil this need. It is not uncommon for Hong Kong Chinese to spend most of their savings in seeking a panacea through Chinese traditional medical practices. This reflects a contemporary Chinese rehabilitation practice which is based on a combination of Western medicine and traditional Chinese medicine (Albrecht & Tang, 1990). The consumer will continue using the services only when the core benefit has been satisfied. The notion that consumers put more emphasis on the medical aspect of the rehabilitation is supported by both the qualitative as well as quantitative findings of the study. First, a large portion of the discussion time was focused on those service attributes related to medical practitioners. Second, two of the three most important perceptual factors (Table 5) are related to the physician's characteristics. However, it is interesting to note that the most important individual attribute in consumers' perception is the competence of the therapist.

Knowing the eight attribute factors and their relative importance, rehabilitation providers in Hong Kong can take actions to improve their competitiveness. For example, advertising can be used to stress a particular attribute on which the rehabilitation provider's service appears to be weak in the mind of the consumers. Alternatively, if the service package does not contain the appropriate mixture of important attributes, then the rehabilitation provider should develop a new service package that incorporates all the important attribute factors. The same process can be applied to managing customer satisfaction. Because satisfaction is the result of the comparison of expectations with perception (Bateson, 1992), consumer perceptions of the service can be improved by stressing in communications how good the service is, relative to competition, on the important attribute factors. Alternatively, care can be taken in communications to ensure the accuracy of customers' expectations on the important attribute factors.

Out of the eight attribute factors, four of the most important factors are related to personnel, two of them related to the system; one related to the physical facility and one related to patient control of the service experience. As distinct from goods where there is a delay between production and consumption, service personnel produce services that are consumed simultaneously by the customers (Lovelock, 1991; Patterson & Marks, 1992). How service personnel conduct themselves in the consumers' presence influences whether consumers buy from the service organization again. It is argued here that for most rehabilitation organizations, personnel are the source of service differentiation as it is often impossible for a rehabilitation organization to differentiate itself from other similar organizations in regard to the core benefit bundle it offers or its delivery system. The second important conclusion from the findings is that human factors should be the prime consideration in formulating the service package.

It should not be surprising to find that the control factor is the least important factor in this study. The study was conducted in Hong Kong where the market in general is relatively undifferentiated and offers limited choices to consumers because the rehabilitation services are largely provided and coordinated by the Government. This may also reflect a deeper Chinese ideology concerning the relation of the individual to the collective. It asserts that new roles will be hammered out in the collective struggle not through individual initiatives of special interest groups. This is why people with disabilities place their care in the trust of the medical collectives (Albrecht & Tang, 1990).

Further research

Not only that the study was conducted in another culture, but the service attributes were derived from a small sample, any generalization of its findings to another population with a different culture must be done with caution. However, as the results are largely consistent with results of similar studies conducted on other populations (Elbeik, 1987; Flexner, McLaughlin & Littlefield, 1977), it may suggest that the findings are generalizable. The author suggests a replication of the study so that the findings would become more relevant and applicable to the local environment. Furthermore, there are three assumptions underpinning this study. First, it is assumed that positive attitudes toward a rehabilitation facility will lead to a purchase outcome. This assumption should at least be tested empirically. Second, it assumes that the measured importance ratings actually reflect salient attributes, that is, those that are actually utilized by consumers in evaluating alternative choices. Researchers (Myers & Alpert 1968; Wilkie & Pessemier 1973) caution us about the use of measured importance ratings in reflecting salient attributes. Third, the population being studied is homogeneous in terms of salient attributes used in determining preferences for rehabilitation facilities. This last assumption has long been challenged, and the pioneering work in the 1930s of Joan Robinson and Edward Chamberlain on an economy theory of imperfect competition and market heterogeneity led to the recognition of the notion of market segmentation by marketeers (Smith, 1956). Studies should be conducted in gaining insights into how the rehabilitation market is segmented.

1 The Hong Kong Workers' Health Center is a non-profit making organization with its major activities in promoting industrial health concepts through community oriented education programs and publication of relevant educational materials. In 1991, the Center had secured a donation amounted to US $150,000 from the Royal Hong Kong Jockey Club Donation Foundation to finance a three-year project involving setting up a rehabilitation service for injured workers.

References

Ajzen, I., & Fishbein, M. (1975). Belief, attitude, intention, and behaviour: an introduction to theory and research. London: Addison-Wesley Publishing Company.

Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and predicting social behaviour. New Jersey: Prentice-Hall, Inc.

Albrecht, G.L., & Tang, X. (1990). Rehabilitation in the People's Republic of China: A reflection of social structure and culture. In Advances in Medical Sociology. Vol. 1, 235-267.

Allport, G. W. (1935). Attitudes. In C. Murchison (Ed.). Handbook of social psychology (3-13). Clark University Press.

Alpert, M. (1971). Identification of determinant attributes: A comparison of methods. Journal of Marketing Research, VIII, 184-91.

Anastasi, A. (1988). Psychological tests (6th ed.). New York: Macmillan Publishing Company.

Bass, F.M., & Talarzyk, W.W. (1972). An attitude model for the study of brand preference. Journal of Marketing Research, November, 456-467.

Bateson, J.E.G. (1992). Managing services marketing (2nd. ed.). Fort Worth: The Dryden Press.

Berkowitz, E.N., & Flexner, W. (1980-81). The market for health care services: Is there a non-traditional consumer? Journal of Health care Marketing, 1 (1), 25-34.

Block, C.E., & Roering, K.J. (1979). Essentials of consumer behavior. Illomois: The Dryden Press.

Boscarino, J., & Steiber, S.R. (1982). Hospital shopping and consumer choice. Journal of Health Care Marketing, 2 (2), 15-23.

Boyd, H.W., Westfall, R., & Stasch, S. (1985). Marketing research text and cases (6th ed.). Illinois: Homewood.

Chisnall, P.M. (1981). Marketing research: Methodological foundations (3rd ed.). London: McGraw-Hill Book Co.

Day, R. L. (1974). Measuring preferences. In R. Ferber (Ed.). Handbook of marketing research (3-101-3-125). New York: McGraw Hill Book Co.

Dillon, W.R., & Goldstein, M. (1984). Multivariate analysis methods and applications. New York: John Wiley & Sons.

Elbeik, M.A. (1986). The measurement of hospital image. Health Marketing Quarterly, 4 (2), 21-33.

Flexner, W.A., McLaughlin, C.P., & Littlefield, J.E. (1977). Discovering what the health care consumer really wants. Health Care Management Review, 2 (4), 43-9.

Goldman, A. E. (1969). The group depth interview. In J.U. McNeal (Ed.). Dimensions of consumer behavior (2nd ed.) (321-336). New York: ACC, Meredith Corporation.

Hasbrook, R. (1981). Editorial: If you have it - flaunt it. Journal of Applied Rehabilitation Counselling, 12(3), 143-145.

Hooley, G.J. (1978). Market segmentation: The segmentation of the UK market for king size cigarettes on the basis of consumer perceptions and product requirements. Unpublished doctoral dissertation, University of Warwick, Coventry.

Hughes, G.D. (1974). The measurement of beliefs and attitudes. In R. Ferber (Ed.). Handbook of marketing research (3-16 - 3-43). New York: McGraw-Hill Book Co.

Kaiser, H.F. (1958). The varimax criterion for analytic rotation in factor analysis. Psychometrika, 23, 187-200.

Klegon, D.A., Kingstrom, P.O., & Gregory, D.D. (1982). Ambulatory health care centers: image and preference among consumers. Journal of Ambulatory Care Management, November, 16-23.

Kotler, P., & Clarke, R. N. (1987). Marketing for health care organizations. New Jersey: Prentice-Hall, Inc.

Lim, J., & Zallocco, R. (1988). Determinant attributes in formulation of attitudes toward four health care systems. Journal of Health Care Marketing, 8 (2), 25-30.

Lovelock, C.H. (1991). Services marketing (2nd. Ed.). London: Prentice-Hall, Inc.

Lowe, A. (1990). Small business survival. Unpublished doctoral dissertation, University of Glasgow, UK.

Matkin, R. (1982). Rehabilitation services offered in the private sector: A pilot study. Journal of Rehabilitation, 48(4), 31-33.

May, V.R. (1985). Current trends in marketing and placing the industrially injured worker. Journal of Rehabilitation, 51(4), 35-7.

Myers, J., & Alpert, M.I. (1968). Determinant buying attitudes: Meaning and measurement. Journal of Marketing, 32, 13-20.

Nadolsky, J.M. (1984). The marketing of rehabilitation services. Journal of Rehabilitation. 50 (3), 4-5,66-7.

Oppenheim, A. N. (1966). Questionnaire design and attitude measurement. London: Heinemann.

Patterson, J.B., & Leach, R. (1987). Client satisfaction in vocational rehabilitation. Journal of Rehabilitation, 53 (2), 40-43.

Patterson, J.B., & Marks, C. (1992). The client as customer: Achieving service quality and customer satisfaction in rehabilitation. Journal of Rehabilitation, 58(4), 16-21.

Rathmell, J.M. (1974). Marketing in the service sector. Massachusetts: Winthrop publishers, Inc.

Riddle, D.I. (1986). Service-Led Growth. Westpoint, CT: Praeger.

Sharpe, L.K., & Granzin, K.L. (1974). Brand attributes that determine purchase. Journal of Advertising Research, April, 39-42.

Smith, S.C., & Clark, M. (1990). Hospital image and the positioning of service centers: An application in market analysis and strategy development. Journal of Health Care Marketing, 10(3), 13-22.

Tabachnick, B. G., & Fidell, L. S. (1989). Using multivariate statistics (2nd ed.). New York: Harper Collins Publishers.

Walker, A. (1988). Rehabilitation: Micro-market or major partner in healthcare's future? Journal of Rehabilitation, 54(1), 19-22.

Wilkie, W.L., & Pessemier, E.A. (1973). Issues in marketing's use of multi-attribute attitude models. Journal of Marketing Research. 10, 428-41.

Woodside, A.G., Nielsen, R.L., Walters, F., & Muller, G.D. (1988). Preference segmentation of health care services: The old-fashioneds, value conscious, affluents, and professional want-it-alls. Journal of Health Care Marketing, 8(2), 14-24.

Horace Ting, The University of Sydney, Cumberland College of Health Sciences, East Street Lidcombe NSW 2141, Australia.