The Use of the General Health Questionnaire in the Australian Defence Force: A Flawed

But Irreplaceable Measure? SARAHCHAPMAN Royal Australian Navy

his article discusses the use of the General Health Questionnaire (GHQ-12) in the Australian Defence T Force (ADF), but comment pertaining to stress meas-

urement and considerations for management are equally applicable to the paramilitary and other high-risk professions. It is concluded that (a) prior to the widespread administration of a questionnaire designed to determine stress levels amongst ADF personnel, the defence research community should attempt to define stress in a conceptual, strategic, and operational fashion; (b) only after an optimal conceptual definition of stress has been operationalised can the defence research community attempt to accurately determine whether (and how) current stress levels are affecting the performance of personnel and begin to impart effective stress-management techniques to those in need; and (c) no one measure can replace the GHQ, which in itself has been inappropriately applied to the ADF population in an attempt to measure the extent to which personnel experience the debilitating effects of physiological and psychological distress.

In addressing the issue of reporting and accountability in stress management within the defence environment, Gilbert (1996) noted that in recent years the Department of Defence has begun to experience the negative impact of increased stress levels amongst employees in the form of increased compensation payments. This finding is concur- rently validated by data indicating abnormally high stress levels amongst uniformed personnel from each of the services (see Grieg, 1997; McIntyre, 1998; Timmins, 1998).’

The 12-item version of the General Health Question- naire (GHQ-12) was originally selected for use in the Australian military by 1 Psychological Research Unit in an attempt to measure levels of work-related or occupational stress amongst members of the Australian Regular Army. It was considered an accurate and rapid measure of psycho- logical distress with robust and well-established psychome- tric properties. Subsequent to its widespread administration in the Australian Defence Force (ADF), the GHQ has come under increased scrutiny. The aim of this review is to deter- mine the appropriateness of the GHQ in providing a suit- able metric for measuring stress levels in ADF personnel and to assess the potential for other instruments to replace this measure.

Psychometric Properties of the Scale The GHQ has primarily been used as a screening device to detect psychatric illness or to estimate the prevalence of psychiatric disorder within a sample of individuals. It may be administered as a 60-, 30-, 28-, or 12-item measure in which subjects respond to a number of state- ments regarding the state of their psychological health (e.g., “Have you recently been feeling sad and gloomy?”). Scoring methods vary, depending on the desired outcome of the administrator. The binary method involves assigning a score of zero to the responses less so than usual and no more than usual, with response options rather more than usual and much more than usual assigned a score of one. The selection of a cut-off score allows determination of psychiatric caseness. The Likert method employs a scoring method in which respondent scores range from 0 (no more than usual) to 3 (much more than usual) for each item. This provides a linear index of the severity of disruption to normal functioning.

Receiver operating characteristic (ROC) analysis is designed to determine which cut-off point (using the binary scoring method) maximises both sensitivity (the proportion of true positives: independently diagnosed cases that the test correctly identifies) and specificity (the proportion of true negatives: non-cases that the test correctly identifies). Sensitivity and specificity have been found to vary considerably, depending on the sample under investigation. Banks et al. (1980) noted that, whilst the binary method discriminates adequately between “cases” and “normals”, the Likert method is more appropriately employed for parametric statistical analyses.2

Piccinelli, Bisoffi, Bon, Cunico, and Tansella (1993) found great similarity in the conventional (binary) scoring method, the Likert scoring method and the revised scoring procedure proposed by Goodchild and Duncan-Jones (C-GHQ)3 in determining caseness, with sensitivity ranging between 71% and 75% and specificity between 73% and 76%. They concluded that no one scoring method is superior to the others (is better able to minimise the number of false negative cases)?

One of the most hotly debated issues surrounding the use of the GHQ is its inability to detect the longevity of respon- dent symptoms (Newrnan, Bland, & Orn, 1988). That is, given that GHQ items are phrased to emphasise recent changes in psychological functioning, the GHQ tends to misclassify individuals with chronic symptoms. Conversely, “individuals experiencing acute stress-related symptoms may

Address for correspondence: LEUT. S.E. Chapman, RAN, Submarine Psychologist, Submarine Training and Systems Centre, HMAS STIRLING, Rockingham WA 6958, Australia.

NOVEMBER 2001 W AUSTRALIAN PSYCHOLOGIST VOLUME 36 NUMBER 3 pp, 244-249

244

THE USE OF THE GEE

be erroneously included as clinical cases because they report sudden changes in symptomatology” (Goyne, 1998, p. 11). Despite the ability of the C-GHQ scoring method to counter this limitation, the GHQ is still subject to criticism for its presumed emphasis on transient symptomatology.

Many researchers have also debated the merits of the different versions of the GHQ. Whilst some researchers caution that respondents are more likely to give affirmative answers to the GHQ-12 (see van Hemert, den Heijer, Vorstenbosch, & Bolk, 1995), others endorse all versions of the GHQ for rapid and accurate determination of general psychiatric status (Lo Bello as cited in Conoley & Impara, 1995). Reynolds (as cited in Conoley & Impara, 1995) discouraged the use of all but the GHQ-60, claiming that the shorter measures fail to withstand psychometric scrutiny. Further to this, he asserted that the GHQ is best (and perhaps solely) suited to providing the general practitioner a clear and objective measure to assist him or her in refer- ring a patient for psychiatric assessment (see also Winefield, Goldney, Winefield, & Tiggemann, 1989).

Gureje and Obikoya (1990) confirmed the use of the GHQ-12 as an initial screening instrument in the detection of psychiatric illness;s however, note with concern the paucity of research examining the validity of the shortest version of the GHQ (but see Banks et al., 1980). Conversely, Goldberg and colleagues (1997) asserted that the GHQ-12, when scored in a binary fashion, provides as accurate an indication of psychiatric morbidity as do the longer versions of the instnunent and the more complex methods of scoring. Additionally, Banks et al. (1980) have confirmed the internal consistency and midimensional factor structure of the GHQ-12 across three different samples.

The brevity of the GHQ-12 makes it attractive for use in busy clinical settings and amongst those of limited liter- acy (Goldberg et al., 1997). The implications for administra- tion of the GHQ-12 within the ADF are clear. In an operational setting, the respondent will be eager to complete the re- quired questionnaire and return to normal duties. It is unlikely that the 12-item measure will tax the attention span of the respondents, and the simple language and low comprehension level of the measure aids in its administra- tion to those members of the ADF whose educational attain- ment may predispose them to reading difficulty?

Given the more sophisticated awareness of stress and its deleterious effects amongst employees in the modem workplace, the face validity of the GHQ and the tendency for subjects to distort their responses is worthy of mention. Schei (1994). in administering the GHQ-12 to a sample of Norwegian Army conscripts, queried the extent to which conscripts had suffered under training or whether the high scores obtained during his research merely reflected a “widespread complaining attitude” (Schei, 1994, p. 43). When this question was put to them, almost a third of respondents agreed with the proposition that they had adopted a negative attitude without due cause. This finding is equally pertinent to the Australian military context, given the ease with which personnel may express dissent on matters that do not directly impede operational effective- ness. Furthermore, it is plausible to suggest that personnel in an operational environment, with an enhanced awareness of diminishing resource allocation, may be highly suspi- cious as to the purpose of the questionnaire and attempt to “fake bad” in order to achieve a favourable financial or material outcome. The GHQ does not presently contain a social desirability scale to determine the likelihood that a respondent is “faking bad”.

Winefield et al. (1989) asserted that the GHQ is resistant to the potential differential effects of sex, socioeconomic

3AL HEALTH QUESTIONNAIRE IN THE AUSTRALIAN DEFENCE FORCE:

status and cultural background on the score of the respondent. Despite this, earlier researchers have claimed that differen- tial response , profiles are indeed observable. For example, McCabe, Thomas, Brazier, and Coleman (1996) found that GHQ-12 scores were affected by age, whilst Hobbs, Ballinger, Greenwood, Martin, and McClure (as cited in McCabe et al., 1996) found that the GHQ was more accurately and reliably applied to men than women. Furthermore, and despite the assertion from Goldberg et al. (1997) that education level did not affect GHQ validity, Mari and Williams (as cited in Goldberg et al., 1997) found that poorly educated respondents were more likely to give positive replies to the questionnaire that were not subsequently validated during interview.

The Utility of the GHQ in the Australian Military Perhaps the most definitive paper addressing the utility of the GHQ in measuring stress levels amongst military personnel is that of Goyne (1998). She noted that the GHQ was first administered to army personnel by way of the annual soldier and officer attitude and opinion surveys, but that research is yet to be undertaken to determine an appropriate cut-off score to determine caseness for an Australian army population.8

Prior to the development of a new tri-service attitude survey, the single services were administering the GHQ to a random, stratified sample of personnel’ on a biennial basis. The question now remains whether to continue administering the GHQ-I2 or to adopt another more expan- sive measure. Currently, there are a number of reasons to caution the use of the GHQ in an Australian military environment. These include: 1. the current lack of consensus in the literature regarding

the most appropriate or accurate scoring method; 2. the lack of substantial normative data within the ADF

that has been validated by concurrent administration of a diagnostic clinical inter vie^;^

3. the lack of research attempting to define the most appro- priate cut-off score within the ADF population to deter- mine psychiatric caseness;

4. that the GHQ measures transient or reactive stress (the GHQ does not provide an indication of the number of personnel suffering chronic disruption to normal psychological functioning, which is arguably a far more useful statistic for defence management purposes);

5 . that, whilst the shorter version of the GHQ is more appropriate for administration in a defence context (for the reasons previously discussed), longer versions of the instrument provide a greater behavioural sample and therefore greater reliability in measurement;

6. that the GHQ screens for psychiatric morbidity, which may or may not be related to work-related stress (deter- mination of which is the primary reason for the adminis- tration of the GHQ by Australian military researchers);

7.that a number of GHQ items may encourage those with a tendency toward hypochondria and psychosomatic illness, or those with a cynical attitude as to the intended purpose of the questionnaire, to distort their responses;

8.that the use of the GHQ to demonstrate the existence of a stress epidemic does not elucidate the nature of the distress experienced by personnel (is the respondent predominantly experiencing depression, or suffering anxiety or some other ailment?) and is therefore of limited utility in assisting those responsible for the emotional wellbeing of personnel to address the problem.

____ NOVEMBER 2001 ‘i’ AUSTRALIAN PSYCHOLOGIST

245

SARAH CHAPMAN

General Issues Surrounding Stress Measurement in the Australian Defence Force In order to more accurately determine the construct validity of the GHQ and to determine its comparative worth as a measure of occupational stress, further investigation of instruments such as the Occupational Stress Inventory (OSI; Osipow & Spokane, 1981) needs to be undertaken. A cursory perusal of the relevant literature suggests that the ability of current measures to determine levels of occupational stress heavily depends on item content. That is, items must correctly and explicitly identify sources of occupational stress to allow respondents to attribute their distress to the appropriate source (see Terry, Nielsen, & Perchard, 1993).

Evidently, the necessary criteria for selecting a measure to accurately determine stress levels amongst ADF person- nel are reliability, validity, and utility (ease of self-adminis- tration, the availability of military or paramilitary data for normative purposes, and the ability of the data to deter- mine areas of weakness andlor need in existing policy and resources). Further to this, the following issues are considered particularly pertinent to the consideration of stress measurement within the ADF 1. Should stress measurement be undertaken in a similar

fashion to the collection of attitude and exit survey data (on an annual or biennial basis)?

2.Should the emphasis in stress measurement be on the concurrent administration of standard health measures (a symptom checklist and a measure of life satisfaction) and the collection of data pertaining to health service utilisation (statistics regarding medical and psychological treatment, suicide rates, compensation claims)?

3. Should the defence research community endorse a stress measure as “the instrument of choice” for researchers examining work-related stress in specific contexts, or should researchers identdy the best means of determin- ing stress levels on a case-by-case basis consistent with their research design and preferred methodology?

4. Is it preferable to measure stress in an objective fashion (visits to health professionals, physiological indices) or to obtain data from the individual based on their appraisal of the number and impact of stressors within their work andor personal life?

5.1s it preferable to obtain a measure of coping behaviour (thereby obtaining a measure of the frequency and perceived success with which individuals employ coping techniques) than to have a respondent provide an assess- ment of the degree to which they experience the debilitat- ing physiological symptoms of stress?

6. Should the ADF focus on occupational stress to the exclu- sion of other forms of stress (which may affect their work performance, such as life satisfaction or financial difficul- ties)?

7. If we are to assume that we can accurately measure stress, which factors best indicate the military respondent’s perceived level of stress and their ability to cope with the stressors they are experiencing?

If we are to focus more exclusively on occupational or work-related stress, the picture becomes only slightly clearer than if we were to focus more broadly on life and job satisfaction. The Office of the Surgeon General (1999) has identified the following factors as indicative of occupational stress:

work characteristics (task design; work quantity, diversity and complexity; resources and equipment; time availabil- ity; deadlines); the physical work environment (noise, light, ventilation, temperature, space, working hours); the nature of the work and its relation to employee temperament, training, skills, and experience; the human environment (organisational structure; management styles, methods, and practice; clarity and perceived fairness of conditions; conflict resolution; direction; communications; training and support; relations with peers and clients; clarity of roles; reasonableness of exceptions and outcome; usefulness of the work; and stability of employment).

Additional factors applicable to a military population are the requirement to operate in an inherently dangerous environ- ment, potential for foreign service, prolonged separation from significant others,1° negative public perceptions of military service, and the impact of these perceptions on morale within the workplace.

Obviously, the ability of any one measure, or indeed any single research design, to measure each of these potential contributors to occupational stress is limited. Further to this assumption, a cursory perusal of the relevant literature suggests that the ability of current measures to determine levels of occupational stress heavily depends on item content. That is, items must correctly and explicitly idenm sources of occupational stress to allow respondents to attribute their distress to the appropriate source (see Terry et al., 1993). It seems that measures such as the Occupational Stress Inventory (Osipow & Spokane, 1998) and the Occupational Stress Indicator (Cooper, Sloan, & Williams, 1988) are the most appropriate measures to achieve this aim.

Other measures that are worthy of further consideration are the Coping in Stressful Situations measure (CISS; Endler & Parker, 1990), the Ways of Coping Questionnaire (WCQ; Folkman & Lazarus, 1988), and the Maslach Burnout Inventory (MBI; Maslach, Jackson, & Leiter, 1996). The CISS and the WCQ are perhaps more appropri- ate for enabling the practitioner to work with clients in strengthening weaknesses in their coping behaviours and in attempting to broaden their repertoire of coping behaviours. However. these measures may also be used to determine the preferred coping styles of operational personnel for whom restrictions in the behaviours accompa- nying their preferred coping strategies are necessarily associated with their employment (cf. Brown, 1999). This will enable health professionals such as military psychologists to fulfil their duty of care and to assist in the development of appropriate stress-management training. In contrast, the MBI has significant potential for administra- tion to personnel in “high risk” (high stress) occupations, such as senior officers and personnel employed in particu- larly demanding operational settings.

The Occupational Stress Inventory The Occupational Stress Inventory (OSI; Osipow & Spokane, 1998) is comprised of three questionnaires that may be administered independently. These are:

the Occupational Roles Questionnaire (for which an example item is “My job requires me to work in seved equally important areas at once”), the Personal Strain Questionnaire (e.g., “My eating habits are erratic”), the Personal Resources Questionnaire (e.g., “I avoid excessive use of alcohol”).

NOVEMBER 2001 Yr AUSTRALIAN PSYCHOLOGIST

246

THE USE OF THE GEF

Reliability and Validity

The OSI contains a significantly greater number of items than the GHQ. Whilst this may be an impediment to ensur- ing its completion by respondents with a limited attention span or low reading ability, it is advantageous insofar as it provides a greater behavioural sample and therefore potentially greater reliability in measurement.

The OSI boasts modest validity, concurrent validity having been demonstrated between the Occupational Roles, Personal Strain and Personal Resources Questionnaires with job satisfaction (Bunda, cited in Conoley & Impara, 1993, burnout, locus of control, absenteeism, and other nonproductive behaviours (Cochran, cited in Conoley & Impara, 1995).

Availability of Normative Data

The manual provides both paramilitary and military norms for the purpose of scoring and interpretation, but the lack of normative data for the instrument in its entirety is cause for concern.

Utility

Arguably, the most significant advantage of the OSI in relation to the GHQ as a potential measure of stress in the ADF is that the former explicitly measures occupational stress, whilst the latter merely provides a generalised index of stress which depends on respondent acknowledgment and experience of physical symptomatology. Specifically, the results obtained from the OSI may provide policy- makers and human resource professionals with a clear indication of deficiencies in organisational resources and practices (“I have the resources I need to get my job done”, “I have divided loyalties on my job”), job design (“I have more than one person telling me what to do”), or personnel selection procedures (’‘I find that I need time to myself to work out my problems”, “I am bored with my work”).

Despite these advantages, it is not recommended that the questionnaire be administered on an annual or biennial basis either within or between services, but rather that it be applied for the purpose of more focused research. This is because the item content is sufficiently specific as to be- come meaningless were the OSI to be applied to personnel employed throughout the ADF and generalisations made as to the nature and extent of occupational stress associated with military employment in Australia.

The Occupational Stress Indicator The Occupational Stress Indicator (OSInd; cooper, Sloan, & Williams, 1998) is a measure designed to “clarify the nature of stress in organisations by identifying sources of stress, intervening factors and the effects of stress on employees” (Conoley & Impara, 1995, p. 620). The OSInd is a 167-item measure comprised of the following subscales:

sources (intrinsic to job, managerial role, relationships with other people, career and achievement, organisational structure and climate, work-home interface), individual characteristics (attitude to living, style of beh- aviour, ambition, broad view Type A, organisational forces, management processes, individual influences, broad view of control), coping (social support, task strategies, logic, home and work relationship, time, involvement), effects (achievement value and growth, job itself, organi- sational processes, personal relationships, broad view of job satisfaction, mental ill health, physical ill health).

?AL HEALTH QUESTIONNAIRE IN THE AUSTRALIAN DEFENCE FORCE:

Witt (cited in Conoley & Impara, 1995) has questioned the scope and veracity of the validation research that has been undertaken and recommends that further research be conducted before the instrument i s endorsed as a measure for use in industry. However, he has no such reservations for its continued use as a research tool, arguing that it has significant potential in that regard.

The initial OSInd scale, “How you feel about your job?’, may be appropriate for administration on an annual or biennial basis in addition to some of the more generic attitude survey items that are currently administered to ADF personnel. Whilst items from the latter instrument attempt to measure similar constructs to those contained within the OSInd scale (such as job satisfaction and task autonomy), they do so with a view to obtaining a more global measure of organisational satisfaction. The OSInd is designed to measure degree to which the respondent is experiencing occupational stress and may prove a useful measure when applied to the ADF for that purpose.

The Maslach Burnout Inventory The Maslach Burnout Inventory (MBI; Maslach et al., 1996) is a 22-item measure comprised of the following three subscales:

emotional exhaustion, (reduced) personal accomplishment, depersonalisation.

As noted previously, the MBI has significant potential for administration to senior officers and personnel working in high stress andlor operational settings. Despite this, its use in conjunction with a measure of occupational stress (such as those outlined above) is not recommended, due to the problem of criterion contamination. Whilst occupa- tional stress and burnout may prove overlapping constructs, concurrent administration of instruments designed to mea- sure each of these constructs will result in contamination of the stress criterion used in the research design. This, in turn, does not enable policy-makers and human resource profes- sionals to apply the research frndings and address deficien- cies in organisational poLicy and resources.

Coping in Stressful Situations measure and the Ways of Coping Questionnaire Some researchers consider it preferable to obtain an indirect measure of stress and stress reactivity by determining the frequency and perceived success with which an individ- ual employs coping techniques. Two of the more promising instruments measuring coping behaviour are the Coping in Stressful Situations measure (CISS; Endler & Parker, 1994) and the Ways of Coping Questionnaire (WCQ; Folkman & Lazarus, 1988). The CISS is a 48-item self- report measure designed to assess the frequency with which the respondent employs task-oriented, emotion-oriented, and avoidance-oriented coping strategies, whilst the WCQ is designed “to identify the thoughts and actions an individ- ual has used to cope with a specific stressful encounter” (Conoley & Impara, 1995, p. 1012). The CISS boasts partic- ularly impressive reliability and validity data.

Despite the conceptual and psychometric virtue of these measures, their use for the purposes of stress measurement on an ADF-wide basis is limited. Neither instrument measures the experience of occupational stress nor identifies its source, necessarily focusing on the stress management of the respondent. These measures are therefore unable to provide policy guidance on stress management. The CISS and WCQ are only recommended for further consideration

NOVEMBER 2001 W AUSTRALIAN PSYCHOLOGIST

247

SARAH CHAPMAN

in the stress measurement debate with respect to specific stress and coping research focused on a discrete population (cf. Brown, 1999). They are deemed to have significant potential in that regard (see Conoley & Irnpara, 1995).

Conclusion The preceding discussion has canvassed a number of quest- ions that are pertinent to the consideration of stress meas- urement in the ADF and the identification of a replacement measure for the GHQ. The GHQ was originally selected for use in the Australian military for the purpose of rapid and robust stress measurement. However, subsequent to widespread endorsement of the GHQ, a considerable body of literature has suggested that it may be appropriate to seek an alternate measure.

The following conclusions are drawn from (an albeit) brief consideration of issues surrounding stress measure- ment in the ADF 1. Researchers tasked with stress measurement in the ADF

should a v o i d criterion contamination by attempting to measure stress using a combination of objective physi- ological data (e.g., health service utilisation), self-report of psychological and/or physiological debilitation, and sources of occupational andor organisational stress (such as lack of human, material, or financial resources and workplace conflict).

2.Annual or biennial stress measurement within the ADF is only recommended for data pertaining to objective indices (such as health service utilisation and compensa- tion claims), but further examination of measures such as the job satisfaction scale of the Occupational Stress Indicator is warranted.

3. It is not advisable that the defence research community endorse a stress measure as the instrument of choice for those wishing to obtain a measure of the extent, preva- lence, and type of stress experienced by ADF personnel because the requirements of a given researcher will vary according to their research design, preferred methodol- ogy, and target population.

4. It is preferable to measure the prevalence and incidence of stress throughout the ADF in an objective fashion (visits to health professionals, physiological indices) than to obtain data from the respondents on preferred coping style and stress management, because this enables policy- makers to focus on addressing deficiencies in organisa- tional policy and resources.

Prior to the widespread administration of a questionnaire designed to determine stress levels amongst ADF personnel, the defence research community should attempt to define stress in a conceptual, strategic, and operational fashion. It is only after an optimal conceptual defdtion of stress has been obtained and validated that we can attempt to accur- ately determine whether (and how) current stress levels are affecting the performance of personnel and begin to impart effective stress-management techniques to those in need.

The four instruments identified in this discussion paper appear able to assist in obtaining conceptual, methodologi- cal, and operational clarification in the stress-measurement debate. However, caution is warranted in relying solely on self-report measures of stress in the absence of more objective data such as health-service utilisation and other physiological indices that impede the operational effective- ness of Australian military personnel.

Endnotes 1 It is feasible to assume that this trend has continued in recent

years, given ADF involvement in a myriad of operational exercises and increased support and assistance with interna- tional peace keeping. Scores obtained using the Likert scoring method are more likely to be normally distributed (i.e., less likely to exhibit skewness and kurtosis) and are therefore more appropriately analysed using paramettic statistical tests. The C-GHQ method entails scoring negatively worded items as 1 rather than 0, thus enhancing detection of those respon- dents suffering a more stable and less transient disruption to psychological functioning.

4 See Gureje & Obikoya, 1990. 5 Concurrent validation was determined using the World Health

Organisation Composite Diagnostic Interview (CDI). 6 Contrary to popular perception, not all ADF applicants are

required to undergo an assessment of literacy and comprehen- sion level. See Tomlinson, Gilks, and Chapman (1998) for an objective analysis of working hours and concomitant stress levels amongst army personnel. Obtaining data from a random, stratSed sample of personnel ensures representativeness of the results and avoids problems pertaining to respondent familiarity with the instrument. This is problematic insofar as it deviates somewhat from the application of the GHQ in the Australian military context (group level analysis and determination of the prevalence of work-related stress), but it remains the best means of validating the measure.

10 This condition of service is most prevalent amongst Navy personnel, but is more recently applicable to army personnel deployed to East Timor.

2

3

7

8

9

References Banks, M.H., Clegg. C.W., Jackson, P.R., Kemp, N.J., Stafford,

E.M., & Wall, T.D. (1980). The use of the General Health Questionnaire as an indicator of mental health in occupational studies. Journal of Occupational Psychology, 53, 187-194.

Brown, L. (1999). Stress and well being in submariners (DSPPR Research Report 2/99). Canberra: Department of Defence.

Conoley, J.C., & Impara, J.C. (1995). (Eds.). The fwelfth mental measurements yearbook Nebraska: Buros Institute of Mental Measurements.

Cooper, C., Sloan, S., & Williams, S. (1988). Occupational Stress Indicator. Melbourne: NFER-Nelson, Australian Council for Educational Research.

Endler, N.S., & Parker, (1990). Coping Invenrory for Stressful Situations (CZSS): Manual. Toronto: Multi-Health Systems.

Folkman, S., & Lazarus, R.S. (1988). Ways of Coping Questionnaire. Melbourne: Australian Council for Educational Research.

Gilbert, J. (1 996). Failures in reporting and accountability in stress management in D e p a m n t of Defence. Unpublished disserta- tion, Defence Management Diploma Program, Australia.

Goldberg, D.P., Gater, R., Sartorius, N., Ustun, T.B., Piccinelli, M., Gureje, O., & Rutter, C. (1997). The validity of two versions of the GHQ in the WHO study of mental illness in general health care. Psychotogical Medicine, 27, 191-197.

Goyne, A. (1998). Use of the GHQ in an Australian Army popula- tion (DSPPR Research Report 4/98). Canberra: Department of Defence.

Goyne, A. (1998). Use of the GHQ in an Australian Army popula- tion (DSPPR Research Report 4/98). Canberra: Department of Defence.

Grieg, J. (1997). Preliminary report on the 1997 RAAF General Attitude Survey: Stress and organisational change. Canberra:

NOVEMBER 2001 W AUSTRALIAN PSYCHOLOGIST

248

THE USE OF THE GENERAL HEALTH QUESTIONNAIRE IN THE AUSTRALIAN DEFENCE FORCE:

Directorate of Psychology - Air Force, Defence Personnel Executive.

Gureje, O., & Obikoya, B. (1990). The GHQ-I2 as a screening tool in a primary care setting. Social Psychiatry and Psychiatric Epidemiology, 25, 276-280.

Maslach, C., Jackson, S.E., & Leiter, M. (1996). Maslach Burnout Inventory. Melbourne: Australian Council for Educational Research.

McCabe, C.J., Thomas, K.J., Brazier, J.E., & Coleman, P. (1996). Measuring the mental health status of a population: A compari- son of the GHQ-12 and the SF36 (MHI-5). British Journal ofpsychiatry, 169, 517-521.

McIntyre, A.T. (1998). Senior army officer attitudes and opinions: An insight into stress and organisational factors. Canberra: Psychological Research Unit, Defence Personnel Executive.

Newman, S.C., Bland, R.C., & Om, H. (1988). A comparison of methods of scoring the General Health Questionnaire. Comprehensive Psychiatry, 29, 402408.

Office of the Surgeon General. (1999, 7-8 June). Stress in the ADF. Senate Legislative Committee Brief.

Osipow, S.H., & Spokane, A. (1998). Occupational Stress Inventory. Odessa, FL: Psychological Assessment Resources.

Piccinelli. M., Bisoffi. G., Bon, M.G., Cunico, L., & Tansella. M. (1993). Validity and test-retest reliability of the Italian

version of the 12-item General Health Questionnaire in general practice: A comparison between scoring methods. Comprehensive Psychiatry, 34, 198-205.

Schei, E. (1994). A strengthening experience? Mental distress during military service: A study of Norwegian army conscripts. Social Psychiatry and Psychiatric Epidemiology, 29, 4W5.

Terry, D.J., Nielsen, M., & Perchard, L. (1993). Effects of work stress on psychological well-being and job satisfaction: The stress-buffering role of social support. Australian Journal

Timmins, P. (1998). The report on the RAN Employee Attitude Survey 1998. (DGNSM Report). Canberra: Department of Defence.

Tomlinson, L., Gilks, D., & Chapman S. (1998). A comparative study of army working hours April 1992 - March I997 (DSPR Research Note 2/98). Canberra: Department of Defence.

van Hemert, A.M., den Heijer, M., Vorstenbosch, M., & Bolk, J.H. (1995). Detecting psychiatric disorders in medical practice using the General Health Questionnaire. Why do cut-off scores vary? Psychological Medicine, 25, 165-170.

Winefield, H.R., Goldney, R.D., Winefield, A.H., & Tiggemann, M. (1989). The General Health Questionnaire: Reliability and validity for Australian youth. Australian and New Zealand Journal of Psychiatq 23,53-58.

of Psychology, 45, 168-175.

- NOVEMBER 2001 W AUSTRALIAN PSYCHOLOGIST

249