Title: A critical commentary on the following research paper:

Kumary, A & Martyn, B. (2008) Stresses reported by UK trainee counselling psychologists. Counselling Psychology Quarterly, March; 21:19-28

The prospect of entering any postgraduate training program can often be intimidating. Not only due to the academic commitment required, but because of the emotional demands and potential financial stranglehold placed on a student. These issues alone can leave trainees in both counselling and related psychological professions vulnerable to stress, which can not only damage the well-being of the student, but lower the overall quality of care experienced by patients when trainees are on placement (Cushway & Tyler 1996; Kumary & Baker 2008).

Kumary & Martyn’s make the simple argument, based loosely around Crushway’s (1992) study of UK clinical psychology trainees, that there are key aspects of training that impact on self-reported stress levels. These included poor supervision, financial costs, childcare, personal therapy and extra supervision. Some of these stressors identified cannot be thought of as essential or necessary aspects of training but this in itself is an area of key debate. (Kumary et al 2008). Other research has also highlighted the same key issues within counselling (Szymanska 2002), but has only looked at one issue in isolation and with this in mind, the present study was an investigation of UK counselling psychology trainees self-reports of their experiences of stress when training. One might go as far to question the rational of any study examining potential stress within such professions considering that having gone through an undergraduate degree already, students are already accustomed to a moderate level of stress and it simply goes with the territory (Cooper & Quick 2003). On the other hand, such a study has never been conducted and may produce compelling results.

Subjects were easy to identify and obtain although only UK counselling trainees who were studying for Part 1 of the BPS diploma were recruited. Are we to assume that this is when stress levels are at an optimal level? Questionnaires were sent out to all institutions. While the general characteristics of the sample are well represented, there was only a 41% return rate. This is good, but not outstanding in comparison to research conducted in similar domains, despite numerous follow-up e-mails and telephone calls (Robertson & Sundstorm 1990). A financial incentive might have improved this return rate, but pre-paid return envelops were provided. However this data was collected in 2003, but not submitted for publication until 2007. Ethically, this delay in analysis and publication produces results that are already four years out of date in an education system that is constantly evolving (Hadley et al 1995).

While such a simple methodology may initially seem sound, on closer inspection, the differences between the training programs of clinical and counselling are only glossed over and there is some evidence to suggest that the disparity between the two hinder a repetition of a seemingly straightforward approach last consulted in 1992 by Cushway.

Aside from the time lapse, it is therefore important to consider the other issues surrounding the modification of a methodology previously used to investigate stress in clinical trainees. Firstly, counselling psychologist training tends to be less scientifically orientated than its clinical counterpart and most NHS posts are only open to Clinical Psychologists (Mayne, Norcross & Sayette 2000). For example, it is generally accepted that counselling psychologists focus more on the therapeutic alliance with clients – having to complete 450 hours of contact by the end of 3rd year training (Hadley et al 1995). While there are considerable similarities between the two disciplines, Norcorss (2000) documents many salient differences including professional activities, theoretical orientations, employment and training settings, graduate admissions and research areas.

The question – what are the differences in training clinical and counselling psychologists? -does not lend itself to an easy answer because psychology can be applied in so many ways. Traditionally, the main difference is in their training and perspective (Mayne et al 2000). It would appear to be an oversight on Kumary et al’s (2008) part to use a similar, modified methodology, previously applied to clinical trainees when the stress causing factors may be quite different. With these differences outlined in more detail, the old methodology would appear to require a more radical modification or adaptation from that used previously.

Two main instruments were used to examine stress within the sample. The Counselling Psychology Trainee Stress Survey (CPTSS) and The General Health Questionnaire (GHQ12) (Wemeke, Goldberg & Yalcin 2000). The CPTSS was developed from Cushway’s (1992) stress survey for clinical psychology trainees with four categories (academic stressors, placement stressors, organizational stressors and personal stressors). What is concerning is the lack of both research confirming the validity of the measure and the small brainstorm session using five trainees under those headings. Furthermore, from their discussion the CPTSS, constructed from 36 items, was only piloted on a further six trainees. This did lead to some changes being made with the authors settling on four descriptive categories slightly different from Cushway’s; academic demands, lack of support systems, placement stressors and personal and professional development. This displays neither convergent or discriminant validity. Finally, despite other more valid forms of questionnaires available measuring stress (for example the Psychological Stress Measure (PMS), this study chose one which was quickly devised from a brief investigation and remains untested in the general population (Lemyre & Tessier 2003; Trovato et al 2006).

The General Health Questionnaire (GHQ12) on the other hand has been specifically validated for use in non-psychotic populations (Wemeke et al 2000). Because it is a shortened, 12-item version of the GHQ, it allows for quick completion, is likely to increase participant response, is quick to code and statistical mistakes also become less likely. For the purposes of this study, it appears to be the ideal choice and has been used to great effect in a large body of pervious work (Winefield, Goldney, Winefield, & Tiggemann 1989; Vaglum & Falkum 1999; Quek, Low, Razack, & Loh 2001). A recent review by Jackson (2007) however, pointed out that the 28 item is usually used because the GHQ28 has been more widely used in other working populations, which allows for better comparisons, but the reliability coefficients have ranged from 0.78 to 0.95 in numerous studies and Jackson concludes (2007, p. 57) that:

‘In using this tool with postgraduate students conducting research in many areas of occupational health, the GHQ rarely fails to provide reliable and effective measures of well-being that usually correlate very highly with other measures of working environments or organizations’

Regardless of how carefully survey data is collected and analyzed, the value of the final result depends on the truthfulness of the respondents’ answers to the questions asked. Over the last twenty years, researchers have debated extensively about the truthfulness of people’s self-reports, and no clear cut conclusion has emerged (Zechmeister, Zechmesiter, & Shaughnessy 2001). If someone is asked whether or not they enjoyed their bath, there is generally no need to question whether this accurately reflects their real feelings. However, in everyday life there are some situations in which researchers should have reason to be suspect. Survey research involves reactive measurement because respondents know that their responses are being recorded. Pressures may be strong for people to respond as they think they ‘should’ rather than what they actually feel or believe (Zechmeister et al 2001). The term used to describe theses pressures is social desirability and in Kumary & Martyn’s study (2008) these issues are present in their entirety (Zechmeister et al 2001). For example, a trainee counselling psychologist’s attitudes towards their own stress and health levels, may be a far cry from their actual stressful behavioural responses.

Both the questioners administered rely solely on self report and this gives rise to some further criticism. The approach is straightforward, but there is a trade-off between allowing for a simple analysis and the complex use of questionnaires in any survey based study. It is a fine balance that is difficult to maintain. Self-report questionnaires are all answered at different times and in different locations by each subject. As a result, the measures are vulnerable to inaccuracies caused by confounding variables. For example, a trainee filling in a stress based measure might have just had a particularly stressful day or experience that will effect their score. They could even have exams in a few weeks. Alternatively, reporting the issue of time-management and stress may be meaningless when the respondent evidently has time to sit down and take part in such a study anyway.

The results from any self report also lack directness. While there is no ideal direct measure of stress, it is possible to get a better indication by measuring some of the physiological effects in the body. For example, stress might be better measured via heart rate, blood pressure, breathing rate, brain waves, muscle tension, skin conductance or temperature (Lemyre et al 2003). While more costly, such a study could be replicated using skin conductance monitors, worn by trainee counselling psychologists and correlate daily activities with any changes. This might produce results documenting what aspects of the course give rise to more stress and allow for re-development and changes to be applied where necessary. A more elementary approach might be to use an electronic pager device which asks every hour, ‘how stressed are you now and what are your currently doing?’

In summary, researchers and clinicians must be careful when adapting clinical tools and methodologies to assess stress. They were designed for pathological disorders and validated using clinical populations and so the statistical distributions are not normal (Trovato et al 2006). As Lemyre et al (2003, p. 1159) state:

‘The concept of stress refers to a set of affective, cognitive, somatic and behavioral manifestations within the range of functional integrity’

Despite this, thirteen items from the CPTSS were identified as being the most stressful issues in the sample population (none came from lack of support), which were split into two groups. The first included practical issues of finding time, funds and suitable placements. One item was also linked with negotiating these three key areas and could have a subsequent impact on their social life. A second group comprised of more general postgraduate issues: academic pressure and professional socialization. In order to determine a basis for the four groupings within the 37 single items of the CPTSS they were employed as four sub-scales (academic, placement, PPD and lack of support). These also gave acceptable levels of reliability. (Kumary et al 2008). The authors also found some good evidence for demographic variants in stress, with significantly higher stress ratings reported by younger participants and lower for those who were older. The GHQ12 results were in two scoring forms – ‘casesness’ and extend of distress with 54 participants identified as ‘cases’ had significantly higher CPTSS scores than the 39 ‘non-cases’. Key findings from Kumary et al (2008, p. 24) included:

‘The higher the stress rated for an aspect of counselling psychology training, the clearer the indicators of psychiatric distress became‘

‘older participants had lower CPTSS ratings – especially on placement issues’

‘men reported lower CPTSS ratings, most notably on academic items’

The support items attracted less attribution in comparison to academic, placement and PPD issues, despite pilot discussions (Kumary et al 2008), suggesting again that the methodology behind this study was flawed from the start. This does to some extent mirror Cushway’s (1992) data in that support was viewed by participants as a resource to ease training-induced stress, and participants viewed it as a resource to be used rather than a cause of stress because it was insufficiently provided. Again, with this knowledge available at the outset, why was the same methodology used?

At this point, one might mention the issue of correlation and how this does not imply causation, but no profile of a stressed student was possible because most of the results were not significant. The authors admit themselves that the data collected is ‘nothing to be proud of’ (2008, p. 25). It is difficult to believe that Krumary et al (2008) did not clearly see the unsophisticated and non-standardized status of the CPTSS as a serious issue before conducting such a study – particularly when compared with more experimental research methods (Lemyre et al 2003). It is possible that the measures used were not sensitive enough to pick up on individual stress differences between participants. The fact remains however, that the fundamental assumptions were wrong and the question remains, do trainees in professions such as clinical and counselling psychology experience more stress than those within the normal population and if so are such emotional demands a critical part of training? Should trainees be exposed to unacceptable stress levels and their apparent resilience used as an assessment criterion of professional suitability? (Hadley & Mitchell 1995)

The basis of this study is not sound enough to warrant any overall generalizations within the target population. The approach was oversimplified at the expense of generalized, poor-quality results. In this sense, the study has contributed little to our knowledge into how trainee counselling psychologists experience stress. The lack of an original approach is a reminder of how academic journals vary in the quality of the research they publish. It is nevertheless important that it was published – to illustrate a methodology that clearly failed and thus prevents further repetition. This is the constant winding road of modern applied psychological research.

References

Cooper, L. C., & Quick, C. J. (2003). The stress and loneliness of success. Counselling Psychology Quarterly, 16, 1-7

Cushway, D. (1992). Stress in clinical psychology trainees. British Journal of Clinical Psychology, 31, 169-179

Cushway, D., & Tyler, P. (1996). Stress in clinical psychologists. British Journal of Clinical Psychologists, 31, 169-179

Goldberg DP, et al. (1978) Manual of the General Health Questionnaire (NFER Publishing, Windsor, England).

Hadley & Mitchell (1995). Counselling Research and Program Evaluation. London: Brooks/Cole Publishing Company

Jackson, C. (2007). The General Health Questionnaire. Occupational Medicine, 57, 79

Kumary, A & Martyn, B. (2008). Stresses reported by UK trainee counselling psychologists. Counselling Psychology Quarterly, 21,19-28

Lemyre, L., & Tessier, R. (2003). Measuring psychological stress – concept, model and measurement instrument in primary care research. Canadian Family Physician, 49, 1159-1160

Mayne, T. J., Norcross, J. C., & Sayette, M. A. (2000). Insider’s guide to graduate programs in clinical and counseling psychology (2000-2001 ed). New York: Guilford.

Norcross C. J. (2000) Clinical Versus Counselling Psychology: What’s the Diff? Eye on Psi Chi, 5 (1), 20-22

Quek, F. K, Low, Y. W., Razack, H. A., & Loh, S. C. (2001). Reliability and validity of the General Health Questionnaire (GHQ-12) among urological patents: A Malaysian study. Psychiatry and Clinical Neurosciences, 55 (5), 509-513

Robertson, M. T., & Sundstrom, E. (1990). Questionnaire design, return rates, and response favorableness in an employee attitude questionnaire. Journal of Applied Psychology, 75 (3), 354-357

Szymanska, K. (2002). Trainee expectations in counselling psychology as compared to the reality of the training experience. Counselling Psychology Review, 17, 22-27

Trovato, M. G., Catalano, D., Martines, G. F., Spadaro, D., DI Corrado, D., Crispi, V., Garufi, G., & Nuovo, S. (2006). Psychological stress measure in type 2 diabetes. European Review for Medical and Pharmacological Sciences, 10, 69-74

Vaglum, P., & Falkum, E. (1999). Self-criticism, dependency and depressive symptoms in a nationwide sample of Norwegian physicians. Journal of Affective Disorders, 52 (1-3), 153-159

Wemeke, U., Goldberg, D., & Yalcin, I. (2000). The stability of the factor structure of the General Health Questionaire. Psychological Medicine, 30, 823-829

Winefield, R. H., Goldney, D. R., Winefield, H. A., & Tiggemann, M. (1989) The General Health Questionnaire: Reliability and Validity For Australian Youth. Australian and New Zealand Journal of Psychiatry, 23 (1), 53-58

Zechmeister, S. J., Zechmesiter, B. E., & Shaughnessy, J. J. (2001). Essentials of Research Methods in Psychology, McGraw-Hill Higher Education