ABSTRACT: The study examined the adequacy of the Chinese version of the General Causality Orientations Scale in measuring motivational deficits among clients with mental illness. The scale was part of a motivational frame of reference designed to evaluate and resolve motivational deficits. The study was carried out in Taiwan, and data were collected at multiple psychiatric facilities. Based on a sample of 353 participants, the translated version was found to be internally consistent and valid. To make the scale more suitable for clinical use, a mixed category was added to the three previously suggested typological categories. It is concluded that the scale can be used in this patient population. Further research is needed to investigate temporal stability and other validity traits of the scale.
Key words: motivation, psychiatric rehabilitation, applied scientific inquiry.
Clients' active participation in treatment activities is highly valued in occupational therapy practice. Through the process of active engagement in activities, clients are helped to develop skills and habits (Burke, 1977). However, occupational therapy clinicians often encounter clients who do not have adequate motivation to participate in treatment programmes (Burke, 1977; Arnsten, 1990). These clients may show a state of passivity or a pattern of inability to be motivated. Without active engagement in treatment activities, clients obtain only minimal benefit from occupational therapy programmes (Kielhofner and Nelson, 1983 ). A lack of motivation is reported as the most difficult and most unsuccessfully overcome obstacle in the occupational therapy clinic (Burke, 1977).
To motivate the client to participate in treatment activities, the importance of determining the client's level of motivation has been addressed in the occupational therapy literature (Arnsten, 1990). Information regarding the client's level of motivation would enable the occupational therapy clinician to identify the treatment environment in which the client would be most likely to succeed and develop a sense of control. This would be achieved by including a set of occupational therapy guidelines, or frame of reference, specifically concerned with the assessment and enhancement of intrinsic motivation. Unfortunately, such frames of reference have yet to be developed (Arnsten, 1990; Carlson, 1997).
The first author recently developed a motivational frame of reference that contains guidelines for evaluating motivational deficits, as well as intervention strategies for resolving such deficits (Wu et al., in press). The target population of this motivational frame of reference is clients with mental illness who are in the process of psychiatric rehabilitation. Although the lack of motivation of clients with mental illness may be associated with various factors, this motivational frame of reference defines motivational deficits as the passive or coercive behaviours that result from cumulative frustration experiences in the past and the client's belief that outcomes of events are uncontrollable.
The motivational frame of reference has a theoretical base composed of learned helplessness theory and self-determination theory (Wu et al., in press). It proposes that occupational therapy clinicians use the General Causality Orientations Scale (GCOS) (Deci and Ryan, 1985a) to determine whether the client has motivational deficits and, if so, the type of deficits. Various intervention strategies are recommended for each type of motivational deficits. The GCOS was developed in the United States, empirically tested on a normal population, and reported as reliable and valid (Deci and Ryan, 1985a). There is no known information about the adequacy of its application in clients with mental illness in the United States or other countries.
The purpose of the study was to investigate whether the GCOS is a reliable and valid instrument for occupational therapy clinicians to evaluate motivational deficits in clients with mental illness. The study was carried out in the researchers' country, Taiwan. Although based on theories developed in the United States, the cultural relevance of the theoretical information had been examined according to the researchers' clinical experiences in Taiwan and the GCOS had been translated into Chinese by the first author.
This study was a part of a continuing effort to develop an adequate frame of reference for evaluating and resolving motivational deficits in clients with mental illness. Results of the study may provide empirical data regarding the use of the GCOS in this patient population, which can provide occupational therapy clinicians with a theory-based and empirically valid instrument to evaluate motivational deficits. Findings of the study may also serve as a basis for future studies on the effectiveness of the intervention strategies in the motivational frame of reference.
Review of the literature
Theories of motivation in psychology
Motivation is the energizer by which human behaviour is started, sustained and stopped (Weiner, 1992). In motivational psychology, the study of human motivation evolved from the machine metaphor, in which human behaviour is reflexive and instinctive, to the Godlike metaphor, in which human behaviour is governed by rational thinking (Weiner, 1992).
The machine metaphor is reflected in earlier motivational theories influenced by Darwin's postulate of human-subhuman continuity. Among the best-known theories are psychoanalytic theory, drive theory and field theory (Weiner, 1992). Psychoanalytic theory suggests that human behaviour is guided by instincts - forces originating inside the body and transmitted to the mind. Drive theory asserts that human behaviour is energized by basic drives, habits and incentives. Field theory states that psychological forces or unmet needs move a person in the environment to attain goals. Within the metaphorical theme of the human as a machine, motivation was measured by the number of repeated trials and calculated by mathematical equations. A person was considered to have no control over his or her motivation. According to Weiner (1992), theories reflecting the machine metaphor contradict everyday observations and, as a result, have been abandoned by motivational psychology.
The Godlike metaphor of human motivation is associated with the belief that humans are created in God's image. Motivational theories included in this metaphor emphasize human cognition (Weiner, 1992). Theories that exemplify the Godlike metaphor include achievement motivation theory, Rotter's social learning theory and attribution theory. In achievement motivation theory, achievement-related behaviour is determined by achievement-related needs, the expectancy of success and failure, and the incentive value of success and failure. Expectancy is the cognitive element in this theory. Rotter's social learning theory stresses learned social behaviour. A person's behaviour is determined by the expectancy of reward and the value of the goal, both of which can be learned through the social contexts of behaviour. In this theory, expectancy of reward is the cognitive element. Attribution theory focuses on how an individual's causal attribution to an event impacts on his or her future behaviour. This causal attribution functions as the cognitive mediator of behaviour. The Godlike metaphor represents the current understanding of human motivation (Weiner, 1992). Motivation is measured by self-reported questionnaires in which human behaviour is considered in social contexts, instead of being restricted to psychological spaces, as viewed by theories encompassing the machine metaphor.
Theories of motivation in occupational therapy
Many occupational therapy clinicians apply the model of human occupation as a conceptual framework to enhance motivation in clients (Arnsten, 1990; Carlson, 1997). The model of human occupation is a conceptual paradigm that depicts a human being as an open system interacting with his or her surrounding environment (Kielhofner and Burke, 1980; Kielhofner, 1995). Among the several areas in this model, motivation is addressed in the volition subsystem within the human system, which governs two hierarchically subordinate subsystems: habituation (formation of habits) and performance (acquisition of skills). Volition functions as an innate, spontaneous urge to explore and master the environment.
In the model of human occupation, the volition subsystem is composed of personal causation, values and interests (Kielhofner and Burke, 1980; Kielhofner, 1995). The construct of personal causation is derived from the work of Burke published in 1977, in which personal causation is referred to as `the initiation by the individual of behavior intended to produce a change in his environment' (Burke, 1977: 256). Burke's construct was, in turn, based on DeCharm's theory of locus of causality (Burke, 1977). The theoretical base for the other components in the volition subsystem, values and interests, is unclear.
In the earlier development of the model of human occupation, locus of control was the variable measured for personal causation (Gusich, 1984; Lederer et al., 1985; Oakley et al., 1985; Smyntek et al., 1985; Barris et al., 1986; Barris et al., 1988; Pizzi, 1990; Bavaro, 1991 ). Such measurement was erroneous. Locus of causality, which served as the theoretical origin of personal causation in the model of human occupation, and locus of control are two different constructs (Deci and Ryan, 1991; Peterson et al., 1993). In motivational psychology, locus of causality belongs to the attribution framework, where cognition is taken into account for human behaviour; whereas locus of control lies in the reinforcement framework, where human behaviour is simply a function of external reinforcement. In the recent development of the model of human occupation, the Volition Questionnaire is used to measure the motivation of clients who have limitations in cognitive or verbal abilities (Chern et al., 1996; de las Heras et al., 1998). The Volition Questionnaire is an assessment instrument that is under development (Chem et al., 1996). For occupational therapy clinicians, a self reported questionnaire on motivation is still lacking.
The motivational frame of reference
Part of the theoretical base of the motivational frame of reference is derived from self-determination theory (Wu et al., in press). This theory stems from Heider's concept of personal causality and its derivatives by DeCharm (Deci and Ryan, 1991 ). At one end of the personal causation continuum, actions are intentionally (personally) caused; at the other, actions are unintentionally (impersonally) caused. In self-determination theory, personal causation lies behind three types of motivation-related behaviours: self-determined behaviour and control-determined behaviour, which are both personally caused; and amotivated behaviour, which is impersonally caused (Deci and Ryan, 1985b; Deci and Ryan, 1987; Deci and Ryan, 1991 ). Self determined behaviour is the overt representation of intrinsic motivation. Although a type of intentional regulation, control-determined behaviour is behaviour that acts on controls in the environment or inside the person, as opposed to the person's own choice. Such behaviour is considered to be extrinsically motivated. Controloriented individuals act simply because they have to, and they usually experience pressure or anxiety. Amotivated behaviour is shown through a state of passivity or disorganization and is associated with the belief that desired goals are unattainable. Such behaviour is usually accompanied by feelings of frustration, fear or depression. Each type of motivation-related behaviour represents the three causality orientations: autonomy, control and impersonal, respectively. In the motivational frame of reference, control-determined behaviour and amotivated behaviour are considered to be the two types of motivational deficit (Wu et al., in press). Readers may refer to the motivational frame of reference for intervention strategies and case examples.
In the motivational frame of reference, clients' level of motivation is measured by the GCOS (Wu et al., in press). The GCOS is a scale that assesses the degree to which a person is oriented in each of the three causality orientations (Deci and Ryan, 1985a). It contains three subscales: ( 1 ) autonomy, measuring self-determined behaviour, (2) control, measuring control-determined behaviour, and (3) impersonal, measuring amotivated behaviour. It has been suggested that by comparing the standardized scores of the three subscales, a person could be categorized into one of the three motivation types: autonomy-oriented, control-oriented and impersonal-oriented (Koestner and Zuckerman, 1994). This categorization method is adopted in the motivational frame of reference for evaluating motivational deficits, and clients who are control-oriented or impersonal-oriented are candidates for the motivational intervention (Wu et al., in press). Because the GCOS is well founded in theory and may be a useful instrument for occupational therapy clinicians to use to evaluate motivational deficits in clients with mental illness, it is necessary to investigate the adequacy of using this instrument in the patient population.
Methods Sample
Participants in the study were selected from seven psychiatric facilities in Taiwan. These facilities were accredited by the Health Department of Taiwan and had given permission to be included in the study. They were in different geographical areas of the country, and clients receiving psychiatric services at each hospital usually came from the surrounding areas. The facilities were either government funded or private. Four of the facilities were psychiatric units in general hospitals, and the other three were psychiatric hospitals. According to the first author's communication with occupational therapists working in the participating hospitals, general hospitals and psychiatric hospitals were different in some aspects. Psychiatric hospitals had greater capacity of patient admission than psychiatric units in general hospitals. The average length of hospital stay tended to be longer in psychiatric hospitals than in general hospitals. And clients receiving psychiatric services at general hospitals usually have better quality of social support than those residing in psychiatric hospitals.
Despite some differences between the two types of facility, each of the participating facilities had some or all of the three types of psychiatric units: day treatment units, rehabilitation units and acute care units. Clients in the day treatment units commuted between their home and the hospital independently or accompanied by a caregiver. Those in the rehabilitation units and acute care units resided in the hospital. Occupational therapy services were regularly provided in each type of unit in the participating facilities. Participants in the study were selected from these psychiatric units.
Potential participants were screened using two criteria: ( 1 ) age 65 or younger, and (2) no obvious cognitive impairments in comprehending the questionnaire used in the study, as determined by the clients' medical records and/or occupational therapists' knowledge about the clients based on clinical observations. Clients who met the study criteria were given the questionnaire, along with an invitation letter from the researchers, explaining the purpose of the study, estimated time for responding to the questionnaire, benefits and possible harms, and assurance of anonymity. Only those who responded to the questionnaire became participants in the study.
The first response in this vignette represents the impersonal orientation, the second, the autonomy orientation, and the third, the control orientation. By summing up the scores for each causality orientation in the 12 vignettes, the GCOS yields three subscores, ranging from 12 to 84. The higher the score, the stronger the orientation.
In the United States, the psychometric properties of the GCOS were investigated by collecting data from 923 undergraduate students and 193 non-students, including mothers, corporate employees, engineers, research scientists and hospitalized cardiac patients (Deci and Ryan, 1985a). The GCOS has been found to be internally consistent (Cronbach's alpha values: autonomy, 0.744; control, 0.694; impersonal, 0.741) and temporally stable (test-retest reliability: autonomy, 0.746; control, 0.711; impersonal, 0.778). The three causality orientations were reported as relatively independent from one another. The subscores of the GCOS were significantly related to the constructs of emotions, attitudes and behaviours in various circumstances and domains. This scale has not been studied in the population of individuals with mental illness.
Procedures
In order to administer the GCOS to individuals with mental illness in Taiwan, the first author translated the instrument into Chinese using ethnographic translation, so that the meaning and cultural content of the source language were maintained in the target language (Hulin et al., 1983). The initial translated version and the original were sent to two Taiwanese doctoral students studying in the United States for a consensus of the translation. These two reviewers did not communicate with one another during the reviewing process. Some minor revisions were made on the reviewers' recommendations.
The revised Chinese version was preliminarily tested by being individually administered to eight clients in one of the participating facilities. These clients were observed while responding to the scale and were individually interviewed immediately after they completed the scale. Seven of the clients finished responding to the scale within 15 minutes. The other client was suffering from involuntary movements because of a recent medication change and had difficulty concentrating on performing tasks, but had no difficulty comprehending the questions. This client was allowed to decide whether to discontinue participation or to take breaks. The client took two breaks and completed the scale within 30 minutes. Although none of the clients experienced any obvious difficulties in responding to the translated scale, they did suggest that a few wordings in the translated version could be simplified for future use with a larger number of individuals with psychiatric illness. They also suggested that the three responses and their corresponding seven-point scales in each vignette should be more conspicuous on the page. Additional revisions were made based on these clients' suggestions.
This Chinese GCOS was then tested at the seven participating facilities. It was administered by occupational therapists working in the participating facilities. Each potential participant was coded with an identification number, and their names were not revealed to the researchers. The scale was administered individually or in a small group, so that the therapist could provide any necessary assistance. When administering the GCOS, the occupational therapists read the invitation letter to the potential participants and answered any questions. The potential participants were then instructed to respond to the scale by practising on the example vignette. They were told to participate in the study by responding to the questionnaire, or to return the blank questionnaire if they refused to participate. The questionnaires were collected by volunteer clients, whether or not they had been responded to. The occupational therapists also assisted in providing the potential participants' demographic data, including diagnoses (DSM-IV), age, gender, education level, date of admission, years since the first onset of psychiatric illness, and whether negative symptoms were observed. It was intended to explore the relationships between these demographic variables and the resultant GCOS scores.
Data analyses
Item analyses (Spector, 1992) were performed to examine item-total correlation for each item and the internal consistency in each subscale. Pearson product-moment correlation was used to determine the relationships among the three subscales. Statistical methods, including t-tests, one-way analysis of variance (ANOVA) and Pearson's product-moment correlation, were conducted to explore the relationships between the GCOS resultant scores and the demographic variables, which were intended to explore evidence of validity of the scale (DeVellis, 1991).
Results
Respondent rates
The Chinese GCOS was given to 427 potential participants who met the research criteria. Eleven (2.6%) of them refused to participate and returned the blank questionnaire. During data entry, 63 (14.8%) questionnaires were found to be invalid for data analyses. Some of these questionnaires were scored without variation (for example, all 7s or all ls), which was very unlikely to happen, and others were not responded to in full. These results may be associated partly with a lack of trust of the study, for most of the clients were not acquainted with the researchers, or with the clients' hesitation to refuse to participate in the study. As a result, questionnaires responded to by 353 (82.7%) clients were counted as valid samples and used in further analyses. Detailed information on respondent rates for the participating facilities is presented in Table 1.
About 60% of the 353 participants were from psychiatric hospitals. Fifty-seven per cent of the participants were inpatients, of whom only 6% were from acute care units, and 43% were outpatients from day treatment units. The percentages of male and female participants were about equal. Eighty-eight per cent of the participants were diagnosed with schizophrenia or its subtypes, 8% with mood disorders and 5% with other mental disorders (including personality disorders, substance-related disorders and adjustment disorders). Because standardized instruments for assessing negative symptoms were not used in any of the participating hospitals, information about negative symptoms was collected according to whether observation of such symptoms was stated in the participants' medical records. As a result, about 55% of the participants were observed with negative symptoms, 21% were not, and information for the other 25% was unavailable. The participants were mostly adults (age range: 13 to 62 years, mean: 33.6) with an average education level of 11 years, and their chronicity of mental illness varied. The demographic data were divided into categorical and continuous variables and are presented in Tables 2 and 3, respectively.
Item analyses
The results of item analyses indicated that four of the 36 items (each from different vignettes) in this Chinese version should be deleted because of low item-total correlation (Spector, 1992). After the deletion, the Cronbach's alpha coefficient for the autonomy subscale was 0.81; control, 0.71, and impersonal, 0.64. Table 4 summarizes the results of item analyses and resultant descriptive statistics of the translated version. The score distribution of the autonomy subscale was negatively skewed (skew/standard error of skewness=-4.18), whereas the other two subscores were normally distributed. The three subscales of the Chinese GCOS were all positively correlated with one another (autonomy-control: 0.66, p<0.001; control-impersonal: 0.47, p<0.001; autonomy-impersonal: 0.44, p<0.001 ).
Resultant scores and demographic variables
Statistical significance in the Chinese GCOS scores was found among some of the categorical demographic variables. For the autonomy subscale, participants from general hospitals scored significantly higher than those from psychiatric hospitals (t=2.36, p=0.019); inpatients scored significantly lower than outpatients (t=-4.09, p<0.001 ); there were statistically significant differences among the three psychiatric units (F=8.34, p<0.001 ), among which participants from the day treatment units scored significantly higher than participants from rehabilitation units (F=16.67, p<0.001); and participants observed with negative symptoms scored significantly lower than those without such symptoms (t=-3.28, p=0.001). For the control subscale, inpatients scored significantly lower than outpatients (t=-2.15, p=0.032); and statistically significant differences were found among the three psychiatric units (F=3.47, p=0.032), among which participants in day treatment units scored significantly higher than those in rehabilitation units (F=5.91, p=0.016).
All three subscores did not differ between the gender groups or among the diagnostic categories. No significant differences were found in the control subscore between types of facility, or whether negative symptoms were observed. For the impersonal subscale, no differences were found in any of the categorical demographic variables. Results of the categorical demographic data and the three subscales are summarized in Tables 5, 6 and 7.
In terms of continuous demographic variables, the autonomy subscore was positively correlated with education level (r=0.124, p=0.021 ), and negatively correlated (that is, an inverse relationship) with length since participants' first psychiatric onset (r=0.116, p=0.035). The impersonal subscore was negatively correlated with the number of past psychiatric hospitalizations (r=-0.168, p=0.008), which was information available in only 248 (70%) of the participants. The control subscore did not correlate significantly with any of the continuous demographic variables. Table 8 presents the correlation matrix of the three subscores of the translated version and the continuous demographic data.
Categorization of motivation types
According to Koestner and Zuckerman (1994), comparing the standardized (z) subscores of the GCOS could categorize respondents into one of the three motivation types: autonomy oriented, control oriented and impersonal oriented. However, results of the study showed that some clients had relatively close z scores in two or all of the three subscales. It seemed arbitrary to categorize these clients as any of the three motivation types, even though Koestner and Zuckerman's criteria seemed to be adequate (E. Deci, personal communication, 2 October 1997).
As the instrument is intended to be used in clinical settings, one concern was that the categorization method would be able to detect motivational deficits in clients and reflect intervention effects. It seemed necessary to separate clients without a salient motivation type from the others, so that it would be possible to examine whether clients without a salient motivation type respond differently to occupational therapy treatment from those with a salient motivation type.
For this purpose, a fourth category was created: the mixed type. To achieve the four-type categorization, it was necessary to select a cut-off value to determine how close for the z scores is too close. Unfortunately, related literature did not provide much help in selecting this cut-off value. Figure 1 shows the percentages of the mixed type in the sample when adopting different cut-off values (in percentile points). It shows that the percentage of the mixed type increases as the cut-off value increases and there is no drastic change in the slope.
By examining the data of the 353 valid samples, 6 percentile points was then selected as the cut-off value because it represents the least value that yielded satisfactory separation results. The 25th percentile of the autonomy subscore, for example, for the autonomy-oriented clients was higher than the 75th percentile for the control-oriented or impersonal-oriented clients. Thus, an additional criterion was added to Koestner and Zuckerman's typological criteria: the highest standardized score must be at least 6 percentile points apart from the other scores. Respondents who did not meet this additional criterion were categorized as the mixed type. With the cut-off value of 6 percentile points, about one-third (34.6%) of the participants were categorized as the mixed type. Descriptive statistics of the four-type categorization are presented in Table 9.
Discussion
Internal consistency reliability
According to the rule of thumb, the minimal acceptable level of consistency reliability is 0.70 (Spector, 1992). The results of the study showed that the autonomy and control subscales in the translated version had good internal consistency, but that the impersonal subscale was slightly below the level of 0.70. Among the 11 retained items in the impersonal subscale, six had an item-total correlation lower than 0.30, which were considered low coefficients and may partly account for the lower internal consistency in this subscale (Spector, 1992). In order to increase internal consistency for the impersonal subscale in the translated version, it seems necessary to further revise or replace items with lower item-total correlation coefficients.
Validity issues
According to self-determination theory (Deci and Ryan, 1987, 1991 ), social contexts in which significant others are involved and which support autonomy are more likely to promote intrinsically motivated, autonomous behaviour. For the two types of psychiatric facility, there was a general impression that clients receiving psychiatric services from general hospitals had a better social support system than those residing in psychiatric hospitals. Family members of the former usually cared about the sick members and their illness, and they were usually devoted to the process of psychiatric care. As the study revealed that participants from general hospitals were more autonomous than those from psychiatric hospitals, this provides evidence of the construct validity of the Chinese GCOS.
Additional evidence of the construct validity of the Chinese GCOS was the relatively stronger autonomy-control association compared with impersonal-autonomy and impersonal-control associations. According to selfdetermination theory, both self-determined behaviour and control-determined behaviour are personally caused, whereas amotivated behaviour is impersonally caused (Deci and Ryan, 1991). Theoretically, it is expected that there would be a stronger association between the autonomy subscore (measuring self-determined behaviour) and the control subscore (measuring control-determined behaviour), and weaker associations between the impersonal subscore (measuring amotivated behaviour) and the other two subscores.
In psychiatry, negative symptoms refer to a collection of phenomena observed among many clients with chronic psychotic disorders, including lack of motivation, social withdrawal, blunted affect and deficits in cognitive functions (Crow, 1980). From this point of view, motivational deficits in clients with mental illness are to some extent associated with negative symptoms. The current study found that participants observed with negative symptoms scored significantly lower on the autonomy subscale than those without negative symptoms. Such a finding provides tentative evidence of the criterion-related validity of the scale. It should be noted that, in this study, negative symptoms were recorded based on informal statements in participants' medical records. For further studies, the relationships between autonomy and negative symptoms need to be investigated by measuring negative symptoms with standardized instruments.
Implications for occupational therapy practice
As revealed in the occupational therapy literature, many occupational therapy clinicians have measured clients' motivation by using evaluation instruments that are inconsitent with the theoretical framework of their practice. Occupational therapy is a health profession that should sustain theoretical bases in its practice (Mosey, 1996). To achieve such a professional standard, occupational therapy clinicians should scrutinize the consistency between the theory and its use in practice, instead of simply including theories in practice.
The results of the study indicated that the GCOS has the potential to be used in occupational therapy practice to measure clients' motivation. The instrument is well founded in the self-determination theory. Although this study investigated the translated version and its clinical applications in Chinese clients, it is likely that the original English version might have similar potential to be used in English-speaking clients. It is suggested that the original version is field-tested among the patient population.
The study results also showed that there was a need to create a mixed type, in addition to the three previously suggested motivation types: autonomy oriented, control oriented and impersonal oriented. With regard to the nature of mental health practice, clients' behaviours are often clinically reasoned in continuous or relative manners, instead of discrete ones. The creation of the mixed category would enable this theory-founded instrument to be more suitable in real situations. In fact, in a later effectiveness study, in which this translated GCOS was used to measure motivation, it was found that the mixed-type clients did not respond as well to a motivational intervention as clients with a salient motivation type (Wu, 1999). Although the phenomenon of the mixed type was not extensively explained in the self-determination theory, findings in the later study support the necessity of treating mixed-type clients separately from the others in practice.
Limitations and recommendations for future studies
In the current study, test-retest reliability of the Chinese GCOS was not investigated. Further research may look at the temporal stability of the translated version. In addition, there was a lack of criterion-related validity of the scale in terms of how it is related to other constructs. Although the study found that participants with negative symptoms scored lower on the autonomy subscale, such criterion-related validity needs to be examined further by using standardized measuring instruments. Furthermore, because of the scope of the study, it was unable to interpret the associations between the GCOS scores and some demographic variables, including education level, length since clients' first psychiatric onset and number of past psychiatric hospitalizations. Future studies may explore the impact of these factors on clients' motivation. Finally, the study results may not represent clients in acute care units, for these clients accounted for only 6% of the study participants. However, given the number of participants and the geographical locations of the sample sources, the study results may be generalized to clients in various locations in Taiwan and other similar psychiatric facilities, in addition to the participating hospitals.
Summary and conclusion
In the current study, four items of the translated GCOS were deleted as a result of item analyses. After the deletion, the autonomy and control subscales had good internal consistency, but the impersonal subscale was slightly below the minimal acceptable standard. The study results provided evidences of construct validity in two instances: ( 1 ) participants whose family members and/or significant others were more involved in the process of psychiatric care showed a higher score in the autonomy subscale (that is, more autonomous), and (2) there was a relatively stronger association between the autonomy and control subscores than other subscore combinations. The study also found that participants with negative symptoms were less autonomous than those without, which may be preliminary evidence of criterion-related validity. To make the GCOS more suitable for clinical use, a mixed category was created in the study to separate clients without a salient motivation type from others, in addition to the three previously proposed motivation types. A cut-off value of 6 percentile points was selected for this purpose. Although the impersonal subscale needs further revision and the temporal stability and criterion-related validity need to be examined, it is concluded that the translated GCOS has the potential to be used in occupational therapy practice to measure motivational deficits among individuals with mental illness.
Acknowledgements
The first author would like to thank Dr Edward Deci, author of the General Causality Orientations Scale, for giving permission to translate and use the scale and for patiently and efficiently answering questions during the study; Shu-Yuan Hu, MA, OTR, and Ting-Yu Anne Chen, PhD, for reviewing the initial translated version of the scale and providing invaluable recommendations; and Professor John Daws, PhD, at the Psychology Department of New York University for his guidance on item analysis and categorization issues of the translated scale. The authors would also like to express sincere gratitude to the administrators, therapists and participating clients at the following facilities for their unconditional support and assistance: the psychiatry departments at National Taiwan University Hospital, Veterans General Hospital Kaohsiung, Lo-Hsu Foundation's Lotung Poh Ai Hospital and the Health Department's Keelung Hospital; and the occupational therapy departments at the Health Department's Pali Mental Hospital, Hung-Chi Psychiatric Institute and Kaohsiung Municipal Kai-Hsien Hospital, Ta-Liao Branch.
The article was part of the first author's dissertation submitted in partial fulfilment of the requirement for the PhD degree at the School of Education at New York University.
References
Awn SCI (1994)- Intrinsic motivation. American Journal of Occupational Thereby 44: 462-::. 3,
Barris R, Dickie V, Kielhofner G ( 1988). A comparison of psychiatric patients and normal subjects based on the model of human occupation. Occupational Therapy Journal of Research 8: 3-23.
Barris R, Kielhofner G, Martin RB, Gelinas I, Klement M, Schultz B (1986). Occupational function and dysfunction in three groups of adolescents. Occupational Therapy Journal of Research 6: 301-17.
Bavaro SM (1991). Occupational therapy and obsessive-compulsive disorder. American Journal of Occupational Therapy 45: 455-8.
Burke JP (1977). A clinical perspective on motivation: Pawn versus origin. American Journal of Occupational Therapy 31: 254-8.
Carlson JL (1997). Evaluating patient motivation in physical disabilities practice settings. American Journal of Occupational Therapy 51: 347-51.
Chem J-S, Kielhofner G, de las Heras CG, Magalhaes LC (1996). The volitional questionnaire: Psychometric development and practical use. American Journal of Occupational Therapy 50: 516-25.
Crow TJ ( 1980). Positive and negative schizophrenic symptoms and the role of dopamine. British Journal of Psychiatry 137: 383-6.
de las Heras CG, Geist R, Kielhofner G (1998). A User's Guide to the Volitional Questionnaire (VQ) (3rd edition). Chicago, IL: The Model of Human Occupation Clearinghouse. Deci EL, Ryan RM (1985a). The General Causality Orientations Scale: Self-determination in
personality. Journal of Research in Personality 19: 109-34.
Deci EL, Ryan RM (1985b). Intrinsic Motivation and Self-determination in Human Behavior. New York: Plenum Press.
Deci EL, Ryan RM (1987). The support of autonomy and the control of behavior. Journal of Personality and Social Psychology 53: 1024-37.
Deci EL, Ryan RM ( 1991 ). A motivational approach to self: Integration in personality. In R Dienstbier (ed.) Nebraska Symposium on Motivation: Vol. 38. Perspectives on Motivation. Lincoln: University of Nebraska Press, pp.237-88.
DeVellis RF ( 1991 ). Scale Development: Theory and Applications (Applied Social Research Methods Series No. 26). Newbury Park, CA: Sage.
Gusich RL (1984). Occupational therapy for chronic pain: A clinical application of the model of human occupation. Occupational Therapy in Mental Health 4(3): 59-73.
Hulin CL, Drasgow F, Parsons CK (1983). Item Response Theory: Application to Psychological Measurement. Monewood, IL: Dow Jones-Irwin.
Kielhofner G (1995). A Model of Human Occupation: Theory and Application (2nd edition). Baltimore, MD: Williams & Wilkins.
Kielhofner G, Burke JP (1980). A model of human occupation, part 1. Conceptual framework and content. American Journal of Occupational Therapy 34: 572-81.
Kielhofner G, Nelson C (1983). A study of patient motivation and cooperation/participation in occupational therapy. Occupational Therapy Journal of Research 3: 35-46.
Koestner R, Zuckerman M (1994). Causality orientations, failure, and achievement. Journal of Personality 62: 321-46.
Lederer JM, Kielhofner G, Watts JH (1985). Values, personal causation and skills of delinquents and nondelinquents. Occupational Therapy in Mental Health 5(2): 59-77.
Mosey AC (1996). Applied Scientific Inquiry in Health Professions: An Epistemological Orientation (Znd edition). Bethesda, MD: American Occupational Therapy Association. Oakley F, Kielhofner G, Barris R (1985). An occupational therapy approach to assessing psy
chiatric patients' adaptive functioning. American Journal of Occupational Therapy 39: 147-54.
Peterson C, Maier SF, Seligman MEP (1993). Learned Helplessness: A Theory for the Age of Personal Control. New York: Oxford University Press.
Pizzi M (1990). The model of human occupation and adults with HIV infection and AIDS. American Journal of Occupational Therapy 44: 257-64.
Smyntek L, Barris R, Kielhofner G (1985). The model of human occupation applied to psychosocially functional and dysfunctional adolescents. Occupational Therapy in Mental Health 5( 1 ): 21-39.
Spector PE (1992). Summated Rating Scale Construction: An Introduction (Quantitative Applications in the Social Sciences No. 82). Newbury Park, CA: Sage.
Weiner B (1986). An Attributional Theory of Motivation and Emotion. New York: SpringerVerlag.
Weiner B (1992). Human Motivation: Metaphors, Theories, and Research. Newbury Park, CA: Sage.
Wu C (1999). Effectiveness of an occupational therapy intervention to facilitate intrinsic motivation of individuals in Taiwan with mental illness who have motivational deficits. Dissertation Abstracts International, 60(12), 2603B. (University Microfilms No. ATT 9935667.)
Wu C, Chen S-P, Grossman J (in press). Facilitating intrinsic motivation in clients with mental illness. Occupational Therapy in Mental Health.
CHIN-YU WU School of Occupational Therapy, National Taiwan University, Taipei, Taiwan, Republic of China
MANN-TSONG HWANG Department of Psychiatry, National Taiwan University Hospital, Taipei, Taiwan, Republic of China
Address correspondence to Chin-yu Wu, School of Occupational Therapy, National Taiwan University, 7 Chung Shan South Road, Taipei, Taiwan 10002, Republic of China. E-mail: [email protected]
Copyright Whurr Publishers Ltd. 2000