This chapter makes a case for the use of a quantitative methodology to investigate workplace bullying in NHS psychologists. Participant selection, data collection and analysis procedures are outlined, followed by a discussion about ethical issues. It will conclude with a brief discussion about some of the methodological issues and dilemmas encountered in the study (INCLUDE IF UNDER POWER).
Before moving on to discuss the methods used, it seems important to first critique the chosen paradigm in order to justify its appropriateness. As such, the following section will briefly consider the contributions of both qualitative and quantitative approaches to the research field. Ultimately, both paradigms offer a tool for understanding the world and the phenomena which exist in it. For the present study, the challenge was to decide which approach would meet the research questions best, and ensure the findings were meaningful to others.
Researchers tend to discuss different research paradigms in terms of their epistemological foundations (e.g. Guba & Lincoln, 1994). In its broadest sense, epistemology refers to the philosophical stance underlying the research. Qualitative research is associated with interpretive epistemology. Essentially, this means it is interested in understanding the meaning behind certain phenomenon, rather than the prevalence of it. The key aspect of qualitative data which makes it distinct from quantitative is that it places the researcher within the context of the situation. One could stipulate that the researcher is immersed in the research. Consequently, predictions are made based on the researchers’ subjective understanding of the data.
Qualitative research has increased in popularity over recent years within the psychology field, resulting in a backlash against quantitative methods. Service user and carer perspectives are now encouraged in psychological research (Kuniavsky, 2003), which might explain the trend towards qualitative methods. Data collection is usually can be achieved through a variety of methods, including collaborative interviews, structured interviews and observations. Data is usually descriptive, and exists in the form of written or spoken words (e.g. transcripts).
Within the context of workplace bullying, however, qualitative methods have attracted a number of criticisms. As Cowie (2003) rightly points out, qualitative methods are not appropriate for investigative studies of workplace bullying. Moreover, data collection methods are typically labour-intensive and time-consuming, for participants and researchers alike. This limitation felt particularly salient for the present study, given that it focused on a busy workforce population. Due to the sensitive nature of the research topic, it was also felt that qualitative data collection methods would not be successful. Face-to-face interviews or focus groups, for instance, would compromise the anonymity of the participants.
The present study was interested in gaining descriptive, comparative data. For this reason, and those just described, qualitative methodology was not chosen.
In contrast, quantitative research is associated with positivist epistemology (Henderson, 1991), which refers to a search for the truth. Some contend the positivist paradigm parallels the traditionalist medical model of care (Higgs & Jones, 2000), as research seeks to be scientific and objective. The positivist researcher aims to keep themselves separate from their research, in order to ensure objectivity. Under the quantitative framework, researchers place emphasis on adherence to a standardised protocol (Kleinbaum et al., 1982). Data can be collected using a variety of methods, for example via surveys or organised experiments. Larger samples are demanded than in qualitative research, in order to ensure the generalisabilty of the findings. Under a positivist paradigm, the rigour of a study is usually assessed by means of reliability and validity (Clark-Carter, 1997).
The majority of workplace bullying research has been conducted using quantitative methodologies based on a positivist belief system. The purpose of the present study was to advance knowledge about the prevalence of workplace bullying in a set of service providers within the NHS – the staff. Prevalence studies of workplace bullying in NHS staff generally use self-report measures (e.g. Quine, 1999; Quine, 2001; Quine, 2002; Steadman, 2009). It is important to note that self-report measures carry a number of limitations within the workplace bullying field, such as under-reporting and response biases (..). The inflexibility of questionnaires also means they only provide limited information regarding the interpersonal processes surrounding workplace bullying (Cowie, 2003).
Nevertheless, they undoubtedly provide the ‘best fit’ for prevalence studies. Self-report measures allow access to a wider sample and therefore give voice to multiple perspectives. This was deemed particularly important for the present study considering the lack of research with the psychologist workforce.
Given the important variables of interest in this study, a positivist, comparative approach was deemed as most appropriate. A quantitative approach was selected for this study because it best addressed the research questions regarding prevalence of bullying within the psychologist workforce and the factors that are associated with it.
The study was a cross sectional postal survey design, in which participants were approached only once, at a single point in time. It investigated the prevalence of workplace bullying (negative acts) and associated occupational health outcomes (i.e. job satisfaction, propensity to leave, well-being, resilience, psychological distress).
Sample and recruitment
The inclusion criteria for participants were as follows (there was no exclusion criteria):
Participants had to be a qualified clinical or counselling psychologist working in the NHS
Participants with a statement of equivalence were accepted
The study adopted an opportunistic sampling method. Potential participants fitting the inclusion criteria were initially contacted by email via their NHS line manager/trust representative (see procedure section for further details). Data was collected by means of a postal survey, therefore there was no face-to-face contact with any of the participants.
An initial calculation by G-power showed a total sample of 110 was necessary in order for the study to be valid.
Sample size and response rates
A total of five NHS trusts participated in the research across the region of South-East England. The total target population wasaˆ¦..Table aˆ¦shows the workforce statistics for each of the participating trusts.
In total, ….psychologists were sent the survey. Of this number, aˆ¦.completed and returned the survey, which represented a response rate of aˆ¦..Response rates for postal surveys vary between …. and …..% depending on attainability and accessibility of the sample population.
As outlined in the participant information sheet (see Appendix..) all participants were given a unique participant number. Participants were advised to quote this number should they wish to withdraw their data at a later date. No participants withdrew their data.
Hypothesis 1: Clinical and Counselling psychologists will experience negative workplace acts.
Hypothesis 2: There will be a different in the levels of psychological distress between those participants who have experienced negative acts and those who have not
Hypothesis 3: There will be a difference in the levels of well-being and resilience between those participants who have experienced negative acts and those who have not
Hypothesis 4: There will be a difference in the levels of job satisfaction and propensity to between those participants who have experienced negative acts and those who have not
The following measures were used for the study were presented to participants as a questionnaire pack. This pack was divided into four sections and included the self-report measures described below. Permission was achieved from all the authors prior to the study. The questionnaires were piloted on two trainee clinical psychologists and two qualified teachers to ascertain how long they would take to complete. Feedback was received on presentation and ease of completion and the questionnaires were amended accordingly. Packs took no longer than 30 minutes to complete.
Section A – Demographic sheet
Section B – Depression Anxiety and Stress Scales (DASS), The Resilience Scale (RS), The Warwick-Edinburgh Mental Well-being Scale (WEMWBS)
Section C – Negative Acts Questionnaire-Revised (NAQ-R)
Section D – Generic Job Satisfaction Scale (GJSS), Propensity to Leave Scale
Section A – Demographic sheet
The demographic sheet was a non-standard measure used to gain a description of the sample. All questions were devised by the researcher and revised in line with guidance from the ethical committee. The questions achieved information about participants’ gender, age (in 10 year bands), professional title (clinical or counselling psychologist), country of qualification, year of qualification, statement of equivalence status, area of speciality and ethnic status (key categories only).
Section B – Occupational health measures
Bullying literature discusses the emotional impact of bullying therefore it was considered important to be aware of the psychological health status of the participants. As previously discussed, strengths-based measures are also lacking in current literature. As such, three measures were included in the second section which investigated psychological distress, well-being and resilience.
The Depression Anxiety Stress Scales (DASS)
The DASS is 42-item standardised measure which includes a set of three scales designed to measure the negative emotional states of depression, anxiety and stress. Each subscale contains 14 items on the scale, and respondents are asked to use a 4-point frequency scale to rate the extent to which they have experienced each state over the past week. It should be noted that the DASS is based on a dimensional concept of psychological disorder. This means there are no direct implications for diagnosis of depression or anxiety in relation to DSM or ICD-10 criteria. Nevertheless, the scores can be sorted into different levels of severity (.i.e. mild, moderate, severe) for each emotional state.
It is well-established that the DASS has good-excellent psychometric properties. The DASS has been shown to have good internal consistency across a range of samples as measured by Cronbach’s alpha (e.g. Brown et al., 1997; Martin et al., 1998, Crawford & Henry, 2003). When compared to two others measures of psychological distress, the DASS was found to have adequate convergent and discriminant validity (Crawford & Henry, 2003). This measure was chosen primarily because it of its strong validity and reliability. In addition, it was thought that psychologists would have had less exposure to the DASS than other scales of psychological distress, for example the Hospital Anxiety and Depression Scale (HADS).
The Resilience Scale (RS)
The Resilience Scale (RS) was developed by Wagnild & Young (1993) and was the first instrument to measure resilience directly. The scale consists of 25 items which reflect five key characteristics of resilience; self-reliance, meaning, equanimity, perseverance and existential aloneness. A detailed description of how each of the scale items relates to these characteristics can be found in Appendix.. Total resilience scores range from 25 (minimum score) to 175 (maximum). Low resilience is indicated by a total score of under 120, whereas moderate-low resilience is indicated by a total score from 121 – 145. Moderate-high, or high resilience is indicated by a total score which is greater than 145. Initially, the scale was tested on a large sample (n=782) of American middle-aged and older adults (Wagnild & Young, 1993). Surveys were sent (n=1500) out using a random sampling method and a response rate of 54% was achieved. Internal consistency of the RS was strong (r=0.91). Since then, the RS has demonstrated high internal consistency reliability, with alpha coefficients ranging from 0.84 to 0.94 (ref). Self-report measures need to demonstrate an internal consistency of at least 0.6 to be considered reliable (Mykletun et al., 2001).
This measure was chosen for the present study primarily because it is a strength-based measure. As discussed in the introduction section, traditional measures included in bullying research have been deficit-based. In addition, the scale is easy to use and quick to complete, which were important given the number of measures included in the questionnaire pack.
Section C – Measure of workplace bullying
The third section included only the primary measure of the present study, the Negative Acts Questionnaire. This achieved information about the prevalence of different negative workplace acts (workplace bullying).
Negative Acts Questionnaire-Revised (NAQ-R)
The Negative Acts Questionnaire (NAQ-R) is now the most frequently-used self-report measure employed in UK workplace bullying research. The original 23-item version was developed by Einarsen et al. in 1994 and included items regarding personal and work-related bullying behaviours. Despite high internal consistency and validity in Norwegian studies, when the NAQ was translated several cultural biases were found. As such, a revised 22-item version was developed (the NAQ-R) which overcame these shortcomings. The psychometric properties of the NAQ-R have been well-established. A recent study (Einarsen et al., 2009) showed the internal reliability of the NAQ-R items to be high when measured by Cronbach’s alpha (0.90). Confirmatory factor analysis revealed three underlying factors: person-orientated bullying, work-related bullying and physical intimidation. All factor loadings exceeded .70, and there were high correlations between the different factors.
Respondents are asked to rate the frequency of different negative workplace acts over the last six months. Some items are task-orientated, (e.g. ‘persistent criticism of your work and effort’) others are person-orientated (e.g. ‘someone withholding information which affects your performance’) and two are related to physical intimidation (e.g. ‘threats of violence or physical abuse or actual abuse’). No reference is made to ‘bullying’ so respondents do not realise what the items are measuring. According to Einarsen and his colleagues (2009) this enables a more objective estimate of bullying behaviours than “self-labelling approaches”. After rating the 22 negative acts, respondents are given a concrete definition of workplace bullying (see introduction section) and asked to confirm whether they have been bullied or not. As such, data derived from the NAQ-R provides information about the frequency of objective bullying behaviours in addition to self-categorisation of bullying. The scale does not provide any information about the severity of the bullying, nor regarding the identity of perpetrators.
It is worth noting that no standardised measures currently exist which measure workplace bullying. At present, the NAQ-R remains to be the most user-friendly measure of bullying at work. It has been used in a number of peer-reviewed published studies and has attracted a plethora of praise in recent articles. As such, it was chosen as the key measure for the present study.
Section D – Job-related measures
The final section included two measures of job-related outcomes; job satisfaction and propensity to leave.
Generic Job Satisfaction Scale
Job satisfaction was assessed using the 10-item Generic Job Satisfaction Scale (GJSS), a self-report non-standard measure developed by Macdonald & MacIntyre (1997). The items assess the strength of job satisfaction using a five-point scale in relation to the respondents’ current position. Interpretation of the GJSS is based on the total score achieved, which falls into one of three categories; low job satisfaction, moderate job satisfaction and high job satisfaction. The GJSS has been validated for use with a range of occupational groups, as is applicable to different age groups (Macdonald & MacIntyre, 1997). Internal consistency was found to be adequate as measured by Cronbach’s alpha (0.77). Positive correlations were found between the scale items and overall happiness. The scale was negatively correlated with job stress, boredom, isolation and risk of illness or injury.
It was chosen as it is easy to use, relatively short and taps aspects of job satisfaction deemed relevant to working in the NHS. For example, it generates information about respondents’ satisfaction with management, as well as their feelings about their supervisor and trust. It also includes a question about job security/insecurity. Given the current financial climate and the fast-changing nature of the NHS, these questions seemed particularly poignant for the present study.
Propensity to Leave Scale
Propensity to Leave was measured using a 4-item self-report measure developed by Bluedorn in 1982. The items assess the likelihood of the person leaving their current position within a given time span (e.g. “How do you rate your chances of quitting in the next three months?”). Responses are given on a seven-point scale ranging from ‘terrible’ (i.e. No chance) (1) to ‘very good’ (i.e. definite chance) (7). A higher score indicates a higher likelihood of propensity to leave.
The Index has an internal consistency
In order to follow ethics committee guidelines, the process for achieving consent from participants was paid particular importance. Potential participants were initially sent an email from a representative in their own trust (e.g. director of therapies) which introduced the research (see Appendix…). Potential participants could opt in by replying to this email with a copy of their work address included. Ethical guidelines stipulated that participants should opt-in to receive the questionnaire pack, rather than opt-out. As such, only participants who consented to be sent the information were sent a pack to the address given in the email. The pack was marked ‘private and confidential’ on the envelope in order to protect it being opened by another member of staff.
The packs included a participant information sheet containing details about the study (see Appendix …), a covering letter (see Appendix …), the questionnaire pack entitled ‘Work Well-Being Survey'(see Appendix…), a consent form (see Appendix…), a freepost addressed envelope and a supportive information sheet (please see paragraph below for details). Once received, participation was voluntary and required written consent. Participants did not have to complete the questionnaire pack and only one mass reminder was sent (from the trust representative) to potential participants via email. Due to the content of the study the researcher wished to avoid making participants feel pressured to participate in the study. All data was collected and securely stored on a computer which only the researcher had access to.
Approval to conduct the research was gained from the relevant NHS Ethics and Research and Development Committees, together with the University Research Ethics Committee. This ensured that the rights and dignity of all participants were protected (see Appendix…for letters of approval). This process involved submitting an application to each committee in accordance with their individual requirements.
It is clear that the research posed some ethical considerations for participants. The sensitive nature of the research may have made some participants feel uncomfortable or distressed after completing the questions, particularly if they had disclosed being a target of bullying. Given these ethical considerations, the following measures were put in place to ensure that participants were best supported.
A telephone hotline number where participants could contact the researcher any week day from 9-5pm to discuss their concerns
Contact details for an independent person (research employee at the University of Surrey) whom participants could contact should they have any concerns about the project which they could not discuss with the researcher
A list of supportive organisations related to workplace bullying and harassment (see Appendix…). This sheet included details of where to find their trust policies on workplace bullying.
Data analysis occurred after data collection.