Multifaceted investigation of occupational wellbeing profiles among guidance counsellors

ABSTRACT

Educational guidance counsellors’ wellbeing at work was studied based on the circumplex model of occupational wellbeing. The main aim was to identify what kind of occupational wellbeing profiles (OWP) representing different levels and combinations of burnout, work engagement, workaholism and job satisfaction existed in this group. Data were collected via e-surveys in two separate samples: Sample 1 in 2019 (n = 211) and Sample 2 baseline in 2020 (n = 343) and follow-up in 2021 (n = 176). Latent profile analysis revealed altogether three OWPs: Satisfied-Engaged (68% in 2019, 53% in 2020, 56% in 2021), Workaholic-Engaged (28%, 42%, 33%) and Burned-out (4%, 5%, 11%). In addition, results concerning Sample 2 with a longitudinal design showed a decline in occupational well-being during the COVID-19.

KEYWORDS:

• Guidance counsellors
• Occupational wellbeing
• Education sector
• COVID-19

Introduction

It has been claimed that “counsellors have a vital role in caring for students’ academic, career, and social/emotional needs, yet [they] often neglect their own needs and wellness” (Winburn et al., 2017, p. 3). This is likely because guidance counsellors’ work, while fulfilling and inspiring, is also very stressful due to multiple work demands, roles and responsibilities such as having large caseloads, many non-counselling duties and lack of proper supervision (Kim & Lambie, 2018; Winburn et al., 2017). Nevertheless, many ethical standards in the field of guidance and counselling include the ideal that counsellors should acknowledge and take care of their own mental work ability as an essential basis for providing high quality services for their clients (ASCA, 2022; CCPA, 2020; IAEVG, 2017).

Hence, the main aim of this study, i.e. the comprehensive investigation of guidance counsellors’ occupational wellbeing considering simultaneously its negative (burnout and workaholism) and positive (work engagement and job satisfaction) indicators, is vital. The present results will provide a more detailed picture of the overall state of guidance counsellors’ wellbeing at work than have earlier studies focusing mostly on the negative and positive indicators of occupational wellbeing in isolation (e.g. Jodoin & Ayers, 2017; Kim & Lambie, 2018; Winburn et al., 2017). Additionally, due to the present study design scrutinising guidance counsellors’ experiences with identical measures in cross-sectional data from 2019 (Sample 1) and different, longitudinal data from 2020 to 2021 (Sample 2), it was possible to ascertain if and how the COVID-19 pandemic impacted guidance counsellors’ occupational wellbeing using Sample 1 as a reference group for Sample 2. The theoretical background of the present study will also provide a framework for guidance counsellors to contemplate their own work-related wellbeing from different perspectives.

Occupational wellbeing profiles and the circumplex model of occupational wellbeing

Occupational wellbeing profiles (OWP) are taken in the present study to represent different intraindividual levels and combinations of work-related wellbeing experiences (for one example, see Mäkikangas et al., 2015). The circumplex model of occupational wellbeing presented by Bakker and Oerlemans (2011) for its part provides the theoretical basis for our investigation of the possible different OWPs among guidance counsellors. The merit of this model is that it approaches occupational wellbeing from a multifaceted perspective and combines the key indicators of employee wellbeing, that is, burnout, work engagement, workaholism and job satisfaction into the same model.

Bakker and Oerlemans (2011) suggest that burnout, work engagement, workaholism and job satisfaction describe the four-fold nature of employee wellbeing experiences. These experiences differ from each other according to two key dimensions of pleasantness, ranging from pleasure to displeasure, and of arousal, ranging from high to low activation. Arousal at work refers to a state of mental attention and stimulus reaction readiness which can range from a state of high activation to a state of low activation as presented in the circumplex model of affect (Russell, 1980) on which the circumplex model of occupational wellbeing is based (Bakker & Oerlemans, 2011).

Hence, burnout reflects emotional states of displeasure and low activation at work (see Bakker & Oerlemans, 2011), and is defined as a persistent, work-related state of ill-being and characterised by exhaustion, cynicism, and professional inadequacy (Salmela-Aro et al., 2010). As an opposite experience, work engagement reflects pleasure and high activation, and is defined as a positive, fulfilling, work-related state of mind and characterised by vigour, dedication, and absorption (Schaufeli et al., 2019). Workaholism, in turn, shares the element of high activation with work engagement. However, the difference between these experiences is that the primary emotional state attached to workaholism is not pleasure but instead displeasure. Hence, workaholism is defined as a strong inner, compulsive drive to work excessively hard instead of working hard due to the joy and fulfilment of work (Schaufeli et al., 2009; Taris et al., 2010). Finally, job satisfaction as an opposite experience to workaholism reflects pleasure and low activation and is defined as individuals’ global positive feelings about their jobs (Spector, 1997).

Recent findings concerning occupational wellbeing profiles

The current understanding is limited concerning the existence and prevalence rates of different types of OWPs among guidance counsellors for at least three reasons. First, when reviewing the research literature, we observed that typically burnout (for review, see Kim & Lambie, 2018) or job/career satisfaction (e.g. Baggerly & Osborn, 2006; Cervoni & Delucia-waack, 2011; Jodoin & Ayers, 2017) were investigated among guidance counsellors whereas focus on work engagement or workaholism has been less common although not completely non-existent (Winburn et al., 2017). Second, more than one occupational wellbeing indicator has been included in the study design rarely. Third, the person-centered approach as a research method (Hofmans et al., 2020; Wang et al., 2013) needed to investigate OWPs was missing as well. In sum, the study of single wellbeing indicators and the utilisation of a variable-centered approach has dominated the research field focusing on guidance counsellors’ occupational wellbeing.

The problem is that the investigation of only a single well-being indicator and a variable-centered approach overlooks the possibility that within the whole sample there may be a varying number of internally homogenous subgroups representing different underlying constellations of burnout, work engagement, workaholism and job satisfaction, in other words, OWPs. However, it is possible to investigate the existence and prevalence of these OWPs when multiple occupational wellbeing indicators are included in the same study design and the data are analysed using a person-centred approach (Hofmans et al., 2020; Wang et al., 2013) as demonstrated below in the studies by Mäkikangas et al. (2015) and Salanova et al. (2014).

Mäkikangas et al. (2015) identified altogether four OWPs among a nationally representative sample of Finnish employees from an age cohort in their 50es: Ordinary (54%; medium work engagement and job satisfaction, low job exhaustion), Engaged (30%; high work engagement and job satisfaction, low job exhaustion), Bored-out (9%; low work engagement and job satisfaction, medium job exhaustion) and Burned-out (7%; high job exhaustion, low work engagement and medium job satisfaction). Salanova et al. (2014) also identified four OWPs among Spanish employees working in services, industry, education, commerce and marketing sectors: Workaholic (30%; high energy and low pleasure), Engaged (29%; high energy and pleasure), 9-to-5 (22%; average energy, high pleasure) and Burned-out (19%; low energy and pleasure). In this latter study, energy comprises job-related high vigour and low exhaustion and pleasure comprises satisfaction with job-related aspects such as one’s tasks, colleagues and organisation.

To compare the identified OWPs between the above reviewed two studies, Engaged and Burned-out resemble each other, as do the Ordinary and 9-to-5 profiles, with average and positive levels of occupational wellbeing and below average levels of ill-health. Consequently, the Bored-out and Workaholic profiles seemed to be sample related. However, it should be noted that these two studies were based on heterogenous samples in terms of occupational status and employment sectors. Thus, it is unclear how well they can be generalised to specific occupational groups, such as the guidance counsellors studied here.

In education sector research there are two studies in which teachers’ wellbeing at work has been studied by using a person-centered approach and including more than one of the key occupational wellbeing indicators. Salmela-Aro et al. (2019) studied Finnish subject teachers (n = 149) working in lower and upper secondary schools and identified two burnout-work engagement profiles: Engaged (30%; high work engagement and low burnout) and Engaged-Burnout (70%; medium work engagement and high burnout). Gillet et al. (2018) studied French high school teachers (n = 312) and identified four workaholism-work engagement profiles: Engaged (22%; low workaholism, high work engagement); Engaged-Workaholic (57%; high workaholism, high work engagement); Workaholic (19%; high workaholism, low work engagement); and Disengaged (2%; low workaholism, low work engagement).

Although we were not able to find a peer-reviewed internationally published study investigating OWPs among guidance counsellors, we have previously conducted a study among 854 guidance counsellors working in lower and upper secondary school, vocational education and training and a university of applied sciences (Rantanen et al., 2020). This peer-reviewed study was published in Finnish only. In this sample three OWPs were identified: Satisfied-Engaged (70%; high work engagement and job satisfaction, low burnout and workaholism); Workaholic-Engaged (25%; high workaholism, medium high burnout, work engagement and job satisfaction); and Burned-out (5%; high burnout and workaholism, low work engagement and job satisfaction).

Together these three studies show that, as illustrated in Figure 1, in comparison to occupational heterogeneous samples (Mäkikangas et al., 2015; Salanova et al., 2014) among teachers and guidance counsellors both favourable (i.e. work engagement and/or job satisfaction) and adverse (i.e. burnout and/or workaholism) indicators of occupational wellbeing can be experienced simultaneously, as seen in the detected contradictory OWPs of Engaged-Burnout (Salmela-Aro et al., 2019), Engaged-Workaholics (Gillet et al., 2018) and Workaholic-Engaged (Rantanen et al., 2020). In turn, as in occupational heterogeneous samples also among teachers and guidance counsellors, OWPs have been observed that show mainly the presence of either favourable or adverse (but not both) occupational wellbeing. This can be seen in the detected favourable OWPs of Engaged and Satisfied-Engaged, and in the adverse OWPs of Workaholic and Burned-out (Gillet et al., 2018; Rantanen et al., 2020; Salmela-Aro et al., 2019).

Figure 1. Occupational wellbeing profiles (OWP) observed among teachers^A, guidance counsellors^B, and in occupational heterogeneous samples (no superscript). Solid outlines refer to OWPs which are in line with the circumplex model of occupational wellbeing (Bakker & Oerlemans, 2011) whereas dashed outlines refer to OWPs that suggest alternative or additional types of OWPs in relation to this model.

Finally, in terms of Figure 1, it seems that among teachers and guidance counsellors favourable or contradictory OWPs are more common, as the prevalence rates range from 22% to 70% for the former and from 25% to 70% for the latter, depending on the sample and occupational wellbeing indicators included in the study. By contrast, adverse OWPs (showing mainly high level of burnout and/or workaholism) have also been presented but seem to be less common, with prevalence rates of 19% (Gillet et al., 2018), 5% (Rantanen et al., 2020) or zero, as this kind of OWP was not observed at all by Salmela-Aro et al. (2019). Yet passive and middle-way OWPs, such as the disengaged profile with a 2% prevalence rate detected only by Gillet et al. (2018) appear to be rarer.

The present study

Earlier research has produced important knowledge about factors that are likely to foster wellbeing or induce illbeing for guidance counsellors at work (e.g. Baggerly & Osborn, 2006; Cervoni & Delucia-waack, 2011; Kim & Lambie, 2018). But this research line does not reveal how guidance counsellors’ overall occupational wellbeing presents itself when studied from a multifaceted perspective (Bakker & Oerlemans, 2011) and using a person-centered approach (Hofmans et al., 2020; Wang et al., 2013). The present study addresses this research gap, aiming to replicate and to considerably extend our earlier study among guidance counsellors that was based on a cross-sectional data set from 2017 and has so far only been published in Finnish (Rantanen et al., 2020). Supporting this and given the summarised results of earlier studies (see Figure 1), the main three OWPs in the education sector in general (that is, not in study-specific terms) seem to be: (1) favourable OWPs showing high work engagement (possibly also high job satisfaction) combined with low burnout and/or workaholism, (2) contradictory OWP showing high work engagement (possibly also high job satisfaction) combined with high burnout and/or workaholism, and (3) adverse OWP showing high burnout and/or workaholism combined with low work engagement and/or job satisfaction (Gillet et al., 2018; Rantanen et al., 2020; Salmela-Aro et al., 2019). In contrast, middle-way or passive OWPs seem less evident in the education sector. Based on these observations, the following first hypothesis was set:

H1: At least three OWPs exist among guidance counsellors. First, a favourable OWP showing mainly high work engagement and/or job satisfaction. Second, a contradictory OWP comprising high work engagement and/or job satisfaction combined with high burnout and/or workaholism. Third, an adverse OWP showing mainly high burnout and/or workaholism.

An additional aim of the present study was to shed light on whether and how the occupational wellbeing of guidance counsellors has changed during the mixed, uncertain and presumably taxing conditions of immediate, hybrid and remote work during the COVID-19 pandemic (Hayden et al., 2021). To investigate this issue, a two-sample study design was utilised. First, in Sample 1, the prevalence rates of the OWPs among Finnish guidance counsellors was investigated before the COVID-19 pandemic with cross-sectional data from spring 2019. Second, in Sample 2 the prevalence rates of the OWPs among Finnish guidance counsellors were investigated during the onset and prolonged COVID-19 pandemic situation with longitudinal data from spring 2020 (baseline) and spring 2021 (follow-up). As both these samples represent the same target population, that is, Finnish guidance counsellors, the earlier contacted Sample 1 provides a pre-COVID-19 cross-sectional reference point for the later Sample 2 with two-wave longitudinal design during COVID-19.

Recent longitudinal studies from the education sector during the COVID-19 pandemic era show a deterioration in employee wellbeing in terms of mental health in Canadian school psychology practitioners (Ritchie et al., 2021), and quality of life in Chilean teachers (Lizana et al., 2021). In addition, work engagement has been detected a declining among Finnish higher education staff, particularly among those whose work engagement was already low at the study baseline (Mäkikangas et al., 2022).

To our knowledge, no studies have been published that would have specifically focused on guidance counsellors, included all four indicators of the circumplex model of occupational wellbeing (Bakker & Oerlemans, 2011), and compared the prevalence rates of OWPs consisting of them before and during the COVID-19 pandemic. Hence, the present study offers valuable new knowledge in this regard and provides one additional view on the question of how employee wellbeing as a multifaceted experience reported by the same occupational group has evolved during the COVID-19 pandemic. Based on the studies presented above, the following hypothesis was set:

H2: In Sample 1, with data collected before the pandemic, the prevalence rates for a favourable OWP (i.e. high work engagement and/or job satisfaction) are lower, and for an adverse OWP (i.e. high burnout and/or workaholism) higher than in Sample 2 with data collected during the COVID-19 pandemic.

Method

Procedure and participants

As described in detail in the following paragraphs, the data for the present study were collected on two independent and separate occasions in collaboration with the Finnish Guidance Counsellors (FGC), which has roughly 1,000 members and is an associate member of the Trade Union of Education in Finland. This study and the content of its e-surveys has been approved by the FGC and anonymous participation has been strictly volunteer at every stage of the study based on informed consent from all participants who were legally competent adults. In these circumstances, both the national and the regulations of the University of Jyväskylä ethical committee exempt this study from an ethical review.

Sample 1. Sample 1 represents here a cross-sectional reference point for Sample 2 and was collected in April – May 2019, when 211 guidance counsellors (84% female and mean age 49 years) responded to an e-survey including the theme of occupational wellbeing. This e-survey was a continuum of a research project “School guidance counsellors at their work” started with the first data collection in 2017 with 854 participants and main results reported earlier in a peer-reviewed national scientific journal (Rantanen et al., 2020). Hence, the 2019 data collection yielded a 25% follow-up response rate when calculated from those that were eligible for the present study. In this study, this meant that they were working in lower secondary school (41%), upper secondary school (20%), both lower and upper secondary school (9%), vocational education and training (23%) or university of applied sciences (7%) as either full-time (81%) guidance counsellors or dividing their worktime between teaching and guidance (19%). Their average weekly working hours including all work-related activities both at school and home was 35.4 (SD = 9.0) and the average amount of pupils/students under their guidance responsibility was 286 (SD = 197).

Sample 2. When the COVID-19 pandemic started, it was decided to launch a new two-wave data collection, “School guidance counsellors at work during the COVID-19 pandemic”, by inviting via email all 1,038 working members of the FGC to participate in an e-survey in April – May 2020 because the previous “School guidance counsellors at their work” research project had already ended. Hence, this new data collection was not based on inviting Sample 1 respondents to participate again and for this reason the participants of Sample 1 and Sample 2 cannot be matched. However, the content (i.e. occupational wellbeing) and data collection timing (i.e. at the end of the school year in spring) was matched between Sample 1 and Sample 2. Of the 431 Sample 2 respondents (42% response rate), 343 (88% female and mean age 51 years) were eligible for the present study. One year later in April – May 2021, the follow-up e-survey was repeated and sent to those 263 matched participants from Sample 2 baseline data who had given permission to be contacted again for this purpose. Of those 197 (75% follow-up response rate) who responded, 176 (87% female and mean age 50 years) were eligible for the present study and their responses form the longitudinal, two-wave COVID-19 experiences part of the present data.

In 2020/2021, of the participants, 32/32% were working in lower secondary school, 23/25% in upper secondary school, 7/6% both in lower and upper secondary school, 28/26% in vocational education and training and 10/11% in university of applied sciences as either full-time (83/83%) guidance counsellors or dividing their worktime between teaching and guidance (17/17%). Their average weekly working hours including all work-related activities both at school and home was 36.9 (SD = 7.3) in 2020 and 37.5 (SD = 6.3) in 2021. The average amount of pupils/students under their guidance/counselling responsibility was 273 (SD = 206) in 2020 and 276 (SD = 173) in 2021.

Measures

Burnout was measured with six items from the Bergen Burnout Inventory, a measure widely used and also well validated in the Finnish context and language (see Feldt et al., 2014; Salmela-Aro et al., 2010). Two items tapped exhaustion (e.g. “I am snowed under with work”), two items cynicism (e.g. “I feel that I have gradually less to give”) and two items inadequacy (e.g. “Honestly I felt more appreciated at work before”) experienced at work. The response scale ranged from 1 = totally disagree to 6 = totally agree, and the Cronbach’s alphas for burnout were .83 (in 2019), .79 (in 2020) and .81 (in 2021).

Work engagement was measured with the three-item, ultra-short Utrecht Work Engagement Scale, a measure widely used and also well validated in the Finnish context and language (see Schaufeli et al., 2019). This scale measures vigour (“At my work, I feel bursting with energy”), dedication (“I am enthusiastic about my job”), and absorption (“I am immersed in my work”) experienced at work. The response scale ranged from 0 = hardly ever to 6 = every day, and the Cronbach’s alphas for work engagement were .84 (in 2019), .78 (in 2020) and .79 (in 2021).

Workaholism was measured with four items from the Dutch Workaholism scale, a measure widely used and also well validated also in the Finnish context and language (see Rantanen et al., 2015; Schaufeli et al., 2009). Two items tapped working excessively (e.g. “I stay busy and keep many irons in the fire”) and two items working compulsively (e.g. “It’s important to me to work hard even when I don’t enjoy what I’m doing”) experienced at work. The response scale ranged from 1 = (almost) never to 4 = (almost) always, and the Cronbach’s alphas for workaholism were .76 (in 2019), .78 (in 2020) and .84 (in 2021).

Job satisfaction was measured with one item to capture overall job satisfaction: “Generally speaking, how satisfied are you with your current job or employment situation?” with a response scale from 1 = totally disagree to 6 = totally agree. A similar item was used by Mäkikangas et al. (2015) for detecting OWPs and following the definition that job satisfaction represents individuals’ global positive feeling about their job (Spector, 1997). The minimum reliability for single-item job satisfaction has been found to be between .45 and .69 (for meta-analysis, see Wanous et al., 1997).

Data analyses

Mplus Version 8.7 software was used for multi-group latent profile analysis (LPA) to investigate what kind of OWPs differing from each other in levels of burnout, work engagement, workaholism and job satisfaction were to be found in Sample 1 providing the pre-COVID-19 ‘reference group’ in 2019 and in Sample 2 forming the COVID-19 ‘onset group’ in 2020 and the ‘prolonged group’ in 2021. LPA is a method for investigating unknown population heterogeneity by identifying clusters of participants with similar response patterns in the observed continuous variables in question (Hofmans et al., 2020; Lubke & Muthén, 2005). In addition, the process of Morin et al. (2016) was utilised to evaluate configural (number of profiles), structural (within-profile means), dispersion (within-profile variability) and distributional (size of the profiles) similarity between the above reference, onset and prolonged groups that were set based on the current study design.

Following the process of Morin et al. (2016), the LPA analyses were started by entering profile indicators simultaneously into the LPA conducted for the COVID-19 reference, onset and prolonged groups separately. In this class enumeration procedure, the aim is to detect an optimal number of profiles occurring within group irrespective of other groups. If class enumeration produces the same number of profiles for each group, “a multiple-group model of configural similarity can then be estimated as a baseline comparison model for the subsequent steps” (see Morin et al., 2016). Here the mean scores for burnout, work engagement, workaholism and job satisfaction were eventually used instead of standardised scores. This was because the LPA results were the same whether mean scores or standardised scores were used as this was inspected. If applicable, mean scores are more transparent and allow better comparison between different studies than standardised scores (Meyer & Morin, 2016).

Table 1. Means and standard deviations for occupational wellbeing indicators and mean-level comparisons between samples and time points.

	Sample 1	Sample 2	Sample 2	t test comparisons in means
	2019, n = 211 M (SD)	2020, n = 343 M (SD)	2021, n = 176 M (SD)	2019 vs. 2020^a	2019 vs. 2021^b	2020 vs. 2021^c
Burnout (1–6)	2.48 (0.92)	2.53 (0.86)	2.65 (0.96)	−0.65 (552), p = .518	−1.78 (385), p = .077	−3.23 (169), p = .001
Work engagement (0–6)	5.40 (0.92)	5.07 (0.83)	4.98 (0.93)	4.34 (552), p = .000	4.45 (385), p = .000	2.74 (169), p = .007
Workaholism (1–4)	2.23 (0.70)	2.52 (0.69)	2.55 (0.77)	−4.78 (552), p = .000	−4.28 (385), p = .000	0.23 (169), p = .817
Job satisfaction (1–6)	5.08 (0.95)	4.74 (0.86)	4.86 (0.92)	4.34 (552), p = .000	2.30 (385), p = .022	2.18 (169), p = .031

^a Independent sample t test between Sample 1 in 2019 and Sample 2 in 2020.

^b Independent sample t test between Sample 1 in 2019 and Sample 2 in 2021.

^c Paired samples t test among those Sample 2 participants who participated both in 2020 and 2021 and were working at the time of both data collections (n = 170).

In deciding the optimal number of latent profiles the following recommended (Celeux & Soromenho, 1996; Hofmans et al., 2020; Morin et al., 2016; Spurk et al., 2020; Tein et al., 2013) fit indices and tests were used: Akaike information criterion (AIC), Bayesian information criterion (BIC), sample adjusted Bayesian information criterion (SABIC), Vuong-Lo – Mendell – Rubin (VLMR) test, Lo – Mendell – Rubin (LMR) test, bootstrap likelihood ratio test (BLRT) and entropy value. Lower AIC, BIC and SABIC values reflect better fit with the data for a given profile solution over another profile solution. The BLRT, VLMR and LMR tests examine whether the k profile solution has a better fit (p < 0.05) than the k-1 profile solution. The profile solution classification quality is in turn assessed via entropy value, which ranges from 0 to 1 with values from .60 to .80 considered acceptable and over .80 preferable. Along with others, Morin et al. (2016) emphasise that it is important to inspect the above listed fit indices and tests in combination and also to give equal importance to the theoretical interpretability and soundness of the profile contents. These guidelines were also utilised in this study.

After the class enumeration procedure, which is a necessary sub-step in configural similarity testing, a configural LPA model was estimated as a first step of Morin et al.’s (2016) sequence of tests of similarity for multiple-group LPA. In this configural model, including Sample 1 (in 2019) and Sample 2 (in 2020 and 2021) data sets as COVID-19 reference, onset and prolonged groups, the profile indicator means and variances across the groups were estimated freely. As a second step, a structural LPA model was estimated in which the profile indicator means across the reference, onset and prolonged groups were constrained equal. As a third step, a dispersion LPA model was estimated in which the profile indicator means plus variances across the reference, onset and prolonged groups were constrained to be equal. As a fourth step, a distributional LPA model was estimated. In this final model, profile probabilities (i.e. profile prevalence rates) across the reference, onset and prolonged groups were constrained to be equal in addition to previous mean and variance equality constraints. The configural, structural, dispersion and distributional LPA models were compared consecutively to each other based on AIC, BIC and SABIC values to determine which constraints were supported and which were not (for more see Morin et al., 2016).

Results

Descriptive results

Table 1 shows the means and standard deviations for burnout, work engagement, workaholism and job satisfaction at each time point as well as the significance of differences in these means between the time points at the whole sample level. Independent sample t-test comparisons showed that work engagement and job satisfaction were reported less and workaholism more by participants in Sample 2 living the COVID-19 pandemic than by participants in Sample 1 reporting these experiences before the pandemic. In addition, within Sample 2, burnout significantly increased over time (mean 2.45 in 2020 and 2.64 in 2021) and a decrease in the mean level of work engagement (from 5.14 to 4.99) and job satisfaction (from 4.82 to 4.66) was observed. However, no time differences were evident in workaholism within Sample 2.

OWPs identified

LPA was used to identify OWPs in Samples 1 and 2 to see whether they replicated the three OWPs detected in our earlier, preliminary study concerning guidance counsellors (Rantanen et al., 2020). As can be seen from Table 2, when the fit indices and tests for LPAs were considered as a whole, a three-profile solution fitted the data best at all time points compared to the other profile solutions. Two out three likelihood ratio tests, that is VMLR and LMR, although not BLRT, indicated that three-profile solution provided better model fit with the data than four-profile solution. Although the p-values for VMLR and LMR were in 2021 borderline significant for better model fit for the three-profile solution over the two-profile solution, the AIC, BIC and SABIC indicated that the three-profile solution surpassed two-profile solution. Furthermore, the entropy values for the three-profile solution were good and the theoretical interpretability and soundness of the three-profile solution were also meaningful (see next paragraph and Figure 2). Hence, this class enumeration procedure supported the configural similarity between the COVID-19 reference, onset and prolonged groups of the present study. As can be further inferred from Table 2, the structural and dispersion similarities were also observed for the three-profile solution as indicated by the lower AIC, BIC and SABIC values for these consecutive models. Instead, distributional similarity was not detected for the three-profile solution, meaning that there were statistically significant differences in the profile sizes, that is, the prevalence rates of the three OWPs detected across the COVID-19 reference, onset and prolonged groups of the present study.

Figure 2. Occupational wellbeing profiles among guidance counsellors in 2019 (n = 211), in 2020 (n = 343) and in 2021 (n = 176). Counts above each bar are observed mean scores except dispersion LPA model means. The latter are estimated sample means from multi-group latent profile analysis model in which the profile indicator means are constrained equal across the time points, that is, the yearly samples ( =  multi-groups) within each profile. Percentages below each diagram refer to the proportion of the whole sample in a given year, that is, prevalence rates of each profile within each year.

Table 2. Fit indices and statistical tests for the estimated Latent Profile Analysis profile solutions and models.

Class enumeration	LL	FP	AIC	BIC	SABIC	VLMR(p)	LMR(p)	BLRT(p)	Entropy	Laten profile proportion %
Time 2019: Reference group, n = 211
1 profile	−1076.94	8	2169.87	2196.69	2171.34	–	–	–	–	100
2 profiles	−987.75	13	2001.49	2045.07	2003.87	.020	.023	.000	.82	76/24
3 profiles	−951.78	18	1939.56	1999.89	1942.86	.038	.041	.000	.86	67/28/4
4 profiles	−931.68	23	1909.35	1986.44	1913.57	.329	.340	.000	.85	61/32/4/3
Time 2020: COVID-19 onset group, n = 343
1 profile	−1649.41	8	3314.81	3345.51	3320.13					100
2 profiles	−1532.22	13	3090.43	3140.32	3099.09	.000	.000	.000	.75	64/36
3 profiles	−1479.11	18	2994.22	3063.30	3006.20	.009	.010	.000	.80	53/42/5
4 profiles	−1451.51	23	2949.02	3037.29	2964.33	.071	.078	.000	.82	51/33/11/4
Time 2021: COVID-19 prolonged group, n = 176
1 profile	−913.71	8	1843.42	1868.79	1843.45					100
2 profiles	−845.10	13	1716.18	1757.39	1716.22	.001	.001	.000	.90	84/16
3 profiles	−813.03	18	1662.07	1719.14	1662.13	.049	.054	.000	.78	56/33/11
4 profiles	−800.65	23	1647.30	1720.22	1647.39	.255	.264	.000	.80	46/40/10/4
Across group similarity
Configural	−4018.09	54	8144.18	8392.20	8220.73	–	–	–	.81	–
Structural (means)	−4056.37	30	8172.74	8310.53	8215.27	–	–	–	.78	–
Dispersion (means and variances)	−4062.08	22	8168.17	8269.21	8199.35	–	–	–	.78	–
Distributional (means, variances, proportions)	−4069.84	20	8179.67	8271.53	8208.03	–	–	–	.79	–

Note. LL = log-likelihood; FP = free parameters; AIC = Akaike information criterion, BIC = Bayesian information criterion, SABIC = Sample-size adjusted Bayesian information criterion, VLMR(p) = p-value for Vuong-Lo – Mendell – Rubin test, LMR(p) = p-value for the adjusted Lo-Mendell-Rubin-test, BLRT(p) = p-value for the bootstrapped likelihood ratio test. Figures in bold face show the chosen three-profile solution for across time point similarity comparisons.

As illustrated in Figure 2, the three OWPs identified differed from each other in the mean levels of work engagement, job satisfaction, burnout and workaholism in a similar fashion in both Sample 1 in 2019 and in Sample 2 in 2020 and 2021 (notice particularly the estimated sample means from the dispersion LPA model in Figure 2) but also in the prevalence rates between the three time points that these two samples together cover. The OWPs detected were labelled according to the study by Rantanen et al. (2020) as the present results replicate these earlier, tentative findings among guidance counsellors. In comparison to Profiles 2 and 3 as well as to whole sample means (see Table 1), participants in Profile 1 experienced the highest work engagement and job satisfaction and the lowest burnout and workaholism. The prevalence rates for this Satisfied-Engaged profile in the present study were 68% in 2019, but only 53% and 56% in 2020 and 2021, respectively.

Participants in Profile 2 for their part reported overall the highest workaholism in comparison to Profiles 1 and 3 (see Figure 2, particularly the estimated sample means from the dispersion LPA model) as well as to whole sample mean (see Table 1). Participants in Profile 2 also experienced more work engagement and job satisfaction and less burnout than did participants in Profile 3, although the opposite was observed in comparison to Profile 1. Moreover, it is noteworthy that with a scale from 1 ‘(almost) never’ to 4 ‘(almost) always’ workaholism reported by participants in this profile was on average 2.9 and with a scale from 0 ‘hardly ever’ to 6 ‘every day’ work engagement reported by participants in this profile was on average 4.5. Hence naming this profile Workaholic-Engaged was not based only comparison to the other OWPs detected and the whole sample mean, but also on the fact that in relation to themselves, the participants in this profile frequently reported experiencing workaholism despite also experiencing work engagement fairly frequently. The prevalence rates for this profile in the present study were 28% in 2019, 42% in 2020 and 33% in 2021.

Finally, the participants in Profile 3 represented the opposite to Profile 1 and reported most burnout, second most workaholism and least work engagement and job satisfaction in comparison to participants in other profiles (see the estimated sample means from the dispersion LPA model in Figure 2 and compare also to whole sample means in Table 1). The prevalence rates for this Burned-out profile were 4% and 5% in 2019 and 2020 respectively, but 11% in 2021.

Finally, as a supplementary, post hoc inspection, the stability of OWP membership was checked using cross-tabulation and χ²-test among those Sample 2 (n = 170 for 2020 and 2021) participants who participated in both data collection rounds. Because cross-tabulation and χ²-test were used, the following findings should be regarded as tentative. Unfortunately, participant sample size under 200 does not allow latent transition analysis, which would have been a preferable method here (Collins & Lanza, 2010; Kam et al., 2016). With these reservations, the participant profile membership across time points seemed to a indicate strong longitudinal stability of the OWPs identified in Sample 2 [χ2 (4) = 69.47, p = .000]. Accordingly, the adjusted residuals were 7.3 for Satisfied-Engaged (n = 78, 46% of the whole longitudinal sample), 5.1 for Workaholic-Engaged (n = 36, 21%) and 4.7 for Burned-out (n = 5, 3%) for participants belonging to the same profile at both time points. Change in profile membership was not common but transitions from one profile to another were observed as follows: from Satisfied-Engaged to Workaholic-Engaged (adj. res. – 4.9, n = 15, 9%) and to Burned-out (adj. res. – 4.2, n = 2, 1%); from Workaholic-Engaged to Satisfied-Engaged (adj. res. – 6.2, n = 19, 11%) and to Burned-out (adj. res. 2.2, n = 12, 7%); and from Burned-out to Satisfied-Engaged (adj. res. – 2.6, n = 1, 1%) and to Workaholic-Engaged (adj. res. – 0.4, n = 2, 1%).

Discussion

The main aim of the present study was to shed light on the question of how guidance counsellors’ overall occupational wellbeing at work in the education sector presents itself when studied from the perspective of a four-fold circumplex model of occupational wellbeing (Bakker & Oerlemans, 2011) and a person-centered approach (Hofmans et al., 2020; Wang et al., 2013). Supporting Hypothesis 1 and replicating our earlier cross-sectional reported findings among Finnish guidance counsellors in 2017 (Rantanen et al., 2020), the same three OWPs were detected for in 2019, 2020 and 2021: (1) Satisfied-Engaged, (2) Workaholic-Engaged, and (3) Burned-out.

Nature and prevalence of OWPs among guidance counsellors vs. others

The content of the Satisfied-Engaged profile we identified corresponds well with previous studies based on teachers (Gillet et al., 2018; Salmela-Aro et al., 2019) as well as with occupationally heterogenous (Mäkikangas et al., 2015; Salanova et al., 2014) samples. However, one notable difference is that the prevalence rate of this OWP has varied between 22% and 30% in these earlier studies, which is much lower than observed prevalence rates for this profile in the present study: 68%, 53% and 56% (in 2019, 2020 and 2021, respectively). The content of the Workaholic-Engaged profile identified here also corresponds fairly well with the Engaged-Workaholic (57%, Gillet et al., 2018) and Engaged-Burnout (70%, Salmela-Aro et al., 2019) profiles observed among teachers but not in occupationally heterogenous samples (Mäkikangas et al., 2015; Salanova et al., 2014). However, again the prevalence rates for this profile differed from teachers being much lower among the guidance counsellors studied here: 28%, 42% and 33% (in 2019, 2020 and 2021, respectively).

In contrast, both the content and the prevalence rates of the Burned-out profile, that is 4%, 5% and 11% in 2019, 2020 and 2021, respectively, corresponds fairly well with the Burned-out profile found in occupationally heterogenous samples (7%, Mäkikangas et al., 2015, p. 19%, Salanova et al., 2014) although not among teachers (Gillet et al., 2018; Salmela-Aro et al., 2019). However, in the latter two studies teachers’ wellbeing at work was not studied fully from the perspective of four-fold circumplex model of occupational wellbeing (Bakker & Oerlemans, 2011), which may explain the difference in results. Hence, the present study together with others (Mäkikangas et al., 2015; Salanova et al., 2014) emphasise the multifaceted investigation of occupational wellbeing in order to recognise employees at the very highest risk. Namely those who lack feelings of pleasure and activation at work and experience only high displeasure combined with very low activation (see adverse OWP2 in Figure 1). The OWPs observed in earlier studies but not in the present study are Ordinary/9-to-5 (22 to 54%), Workaholic (19 to 30%) and Bored-out/Disengaged (2 to 9%) (cf. Gillet et al., 2018; Mäkikangas et al., 2015; Salanova et al., 2014).

Based on the above research findings comparisons it seems that guidance counsellors might on average feel better at work than teachers or the working population in general if one considers only the prevalence rates for the Satisfied-Engaged profile. These were high in the present study. Supporting the present findings Yeşilyaprak and Boysan (2015) have also reported that school counsellors experience greater levels of job satisfaction than do administrators and office staff. One explanation for this might be the nature of guidance counsellors’ work. Under reasonable working conditions it enables quite well the four important pathways to meaningful work (e.g. Martela & Riekki, 2018): work autonomy (e.g. possibility to affect one’s work methods and schedule), competence (e.g. mastering both lessons, group and individual counselling), relatedness (e.g. collaboration and networking both within and outside one’s school/college) and beneficence (e.g. making a positive contribution for pupils/students and society) (cf. Baggerly & Osborn, 2006; Lairio & Nissilä, 2002; Winburn et al., 2017). Meaningful work in turn has been reported to have very strong correlations (> .70) with work engagement, work commitment and job satisfaction (meta-analysis by Allan et al., 2019). Nevertheless, more research, and preferably with cross-country comparisons, is needed before it can be concluded that guidance counsellors are a fortunate occupational group in relation to others in the education sector or in other industries.

On the other hand, as among teachers (Gillet et al., 2018; Salmela-Aro et al., 2019) also among guidance counsellors the absence of a middle-way OWP (see Figure 1), that is, the Ordinary or 9-to-5 profile (Mäkikangas et al., 2015; Salanova et al., 2014) and the presence of Workaholic-Engaged profile may perhaps predispose them to mild or severe job exhaustion, even burnout. This might happen if work engagement in combination with workaholism turns into overcommitment and an inability to keep healthy temporal borders between work and nonwork, and to detach oneself psychologically from work-related duties and particularly from troublesome issues (see also Winburn et al., 2017). These factors are among key determinants for recovery from normal work strain which support one’s occupational wellbeing along with reasonable workload and psychosocial working conditions (Kinnunen et al., 2016; Sonnentag et al., 2010).

Theoretically, work engagement and workaholism have been presented as separate experiences, as the former refers to joy at work that drives one to work with high vigour and dedication, and the latter refers to an inner obsession that forces one to work hard even without joy (Taris et al., 2010). Hence, it is interesting that earlier among teachers (Gillet et al., 2018) and now also among guidance counsellors, this kind of contradictory OWP where work engagement and workaholism are both experienced simultaneously has been detected with more than marginal prevalence rates (here approximately a third of the participants and 57% in Gillet et al., 2018). When this contradictory OWP is considered from the perspective of the circumplex model of occupational wellbeing (Bakker & Oerlemans, 2011), it can be seen that high investment, energy and arousal at work is the component that is common for work engagement and workaholism, but pleasure should not be. Namely, according to the circumplex model pleasure is claimed to be high in the experience of work engagement and low in the experience of workaholism.

However, the emergence of a contradictory OWP with high workaholism and work engagement in our study with two separate samples and also in another study in the education sector (Gillet et al., 2018), may suggest a broader understanding and manifestation of pleasure at work. Accordingly, a very tentative and cautious conclusion could be that among some employees, and maybe particularly among some of those working in the education sector, workaholism may be associated with experience of high rather than low pleasure but this pleasure is perhaps more obsessive, compulsive and energy-draining in nature. Conversely, pleasure relating to work engagement without concurrent experience of workaholism is perhaps more self-fulfilling and energy-charging in nature (see also a recent discussion on this issue by Cook & Gilin, 2023, in their study of concurrent experiences of work engagement and workaholism). Furthermore, it is quite likely that employees reporting a combination of high work engagement and workaholism do not experience these feelings at the very same moment at their work but instead these experiences may fluctuate and occur interchangeably, for example, during the working week, month and year. Hence, more research with both short-term diary and long-term, multi-wave study designs and occupationally large and diverse samples is definitely warranted to see whether this OWP is truly a double-edged sword in terms of employee wellbeing or not, and whether its presence is restricted to the education sector or not.

Prevalence of OWPs among guidance counsellors before and during COVID-19 pandemic

An additional aim of the present study was to examine whether guidance counsellors working in the education sector experienced their occupational wellbeing differently from before, when, together with others, they were confronted with the rapidly changing, taxing and exceptional circumstances in work and personal life due to the COVID-19 pandemic (Hayden et al., 2021). Supporting hypothesis 2, we found tentative evidence that the pandemic possibly had a negative effect on guidance counsellors’ wellbeing at work. The prevalence rate for the Satisfied-Engaged profile was 68% in our reference Sample 1, reporting their experiences before the pandemic, but these rates were 53% and 56% in our COVID-19 Sample 2, reporting their experiences during pandemic in 2020 and 2021 respectively. In addition, the prevalence rate for the Burned-out profile was found to be 11% in 2021, whereas it was respectively 4% and 5% in 2019 and 2020. These findings are in line with earlier longitudinal studies from the education sector showing a decrease in work engagement (Mäkikangas et al., 2022), mental health (Ritchie et al., 2021) and quality of life (Lizana et al., 2021) during the pandemic. What is interesting in the present results is that Workaholic-Engaged profile shows a peak of 42% in 2020 in comparison to 28% in 2019 and 33% in 2021.

Taken together, the pattern of the present results based on our reference Sample 1 and our COVID-19 Sample 2 would suggest that, among guidance counsellors, the presumably rapid and perhaps taxing changes that the pandemic caused to their work initially induced workaholic tendencies, that is, high inner pressure to work excessively, hard, and out of a sense of duty rather than of joy. Only after the pandemic became prolonged did more severe feelings of job exhaustion and burnout symptoms increase, together with a decrease in feelings of vigour, dedication, and absorption. Yet at the same time over half of the studied guidance counsellors belonged to the Satisfied-Engaged profile even at the height of the pandemic as seen in Sample 2. This may relate to the fact that guidance counsellors’ meaningful work as well as perhaps their work-related personal strengths may have given many of them the resources and tools to confront challenges and misfortunes in working life.

Among the Sample 2 participants of the present study, it was also possible to tentatively scrutinise belongingness to a same OWP versus moving from one OWP to another over time. The tentative findings for this show that the stability of the OWPs identified was very high: 70% of the guidance counsellors studied remained in the same OWP between 2020 and 2021. Hence, transitions from one OWP profile to another were rare. Mostly these concerned transitions from the Satisfied-Engaged profile to the Workaholic-Engaged profile (9%) and vice versa (11%), and from the Workaholic-Engaged profile to the Burned-out profile (7%). These findings support the already mentioned need to study the role and replication of the Workaholic-Engaged profile in more detail.

In addition, it will be important to follow-up on the occupational wellbeing of employees in all industries to see what happens to it after the COVID-19 pandemic has subsided. Will, for example, the prevalence rates of favourable OWPs rise and adverse OWPs decline and if so, to what kind of working life conditions and processes are these changes related? Will organisations and employees return to old and proven or new and good-observed ways of working or find a balance between them? In the context of guidance and counselling one particularly timely theme relating to this is the purposeful and effective use of information and communication technology in conjunction with face-to-face practices (e.g. Dores et al., 2020; Sampson et al., 2019), which may also have influence on occupational wellbeing via, for example, work-related flexibility or task variety.

Limitations, conclusions and practical implications

The first limitation of the present study is that the observed OWPs and their prevalence rates are based on self-evaluative survey data instead of, for example, on data triangulation from co-workers, supervisors, or occupational health care professionals. Second, the sample sizes of the present study are not particularly high. This may have affected the content and number of OWPs detected although they are in line with our earlier tentative findings with a larger sample of 854 Finnish guidance counsellors (Rantanen et al., 2020). Third, using latent transition analysis (Collins & Lanza, 2010; Kam et al., 2016) as a preferred method to examine the longitudinal stability vs. changes in OWP membership was not possible due to the small sample size. Fourth, our study design would have been stronger in terms of investigating guidance counsellors’ occupational wellbeing across the pre, onset and during COVID-19 situation if we had opted for a true three-wave longitudinal design instead our cross-sectional Sample 1 (in 2019) and separate longitudinal Sample 2 (in 2020 and 2021). Finally, in future research, it would be advisable to scrutinise the interrelations between the key indicators of OWPs (i.e. burnout, work engagement, workaholism and job satisfaction) by incorporating both variable – and person-centered approaches into the same study design (for a more on this issue see Morin et al., 2017). This procedure might reveal even better than variable-centered or person-centered approach alone how the different indicators of occupational wellbeing combinate and intertwine with each other.

Keeping the above listed limitations in mind, in the present study the multifaceted phenomenon of occupational wellbeing among guidance counsellors was studied relatively comprehensively. With multiple time points in two samples, this study also offers a unique view on guidance counsellors’ experiences of wellbeing at work before, at the start of and during the COVID-19 pandemic. In addition, the present study gives strong support to the view presented in Figure 1 that diverse favourable and adverse OWPs exist among employees from different industries. However, the presence of contrary OWPs as well as the absence of middle-way and passive OWPs seems somewhat unique to the education sector.

One practical implication of the present study is that Figure 1 can be used in working life and career counselling as a basis for self-reflection and discussions among employees in all industries (including guidance counsellors themselves) when the aim is to identify one’s own state of occupational wellbeing, recognise its different manifestations, and analyse factors that contribute to it in one’s unique work-life situation. This can be done when the citations from Figure 1 are omitted, psychoeducation is given to counselees to understand the basic tenets of the circumplex model of occupational wellbeing as well as possible alternative types of OWPs, and reflective questions are offered to increase counselees self-knowledge. This kind of client version of the circumplex model of occupational wellbeing answers, in part, the call for better integration of theory, research, and practice in the field of guidance and counselling (e.g. Sampson et al., 2014). Another practical and everyday implication following from the present results is the importance of charting not only burnout “talk” and symptoms among guidance counsellors in everyday supervisor work as well as in more preventive HR – and occupational health encounters. Attention should also be paid to workaholism-related tendencies to work obsessively and excessively. These unhealthy tendencies sometimes go hand in hand with generally sought and highly valued work engagement.

Acknowledgements

We thank all the participants for their time, guidance counsellors Jenni Jussila and Taru Leikas for their effort in the preliminary preparations of this study, and anonymous reviewers for their exceptionally perceptive and constructive comments.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The authors have full control over the data reported in this text and agree to make the materials and data available to the journal and other researchers provided that their requests do not violate the European Data Protection Regulation statutes.

Additional information

Notes on contributors

Johanna Rantanen

Johanna Rantanen, PhD and docent (adjunct professor) in Psychology, works as a senior lecturer in the Department of Psychology, University of Jyväskylä, Finland.

Anne Mäkikangas

Anne Mäkikangas, PhD and docent (adjunct professor) in Psychology, works as an associate professor (tenure track) and leader of the Work Research Centre in the Faculty of Social Sciences, Tampere University, Finland.

Sauli Puukari

Sauli Puukari, PhD and docent (adjunct professor) in Education, worked at the time of the preparation of this study as a senior lecturer in the Department of Teacher Education, University of Jyväskylä, Finland.

Seija Koskela

Seija Koskela, PhD in Education, works as a senior lecturer in the School of Professional Teacher Education, Jamk University of Applied Sciences, Finland.

Anu Raudasoja

Anu Raudasoja, PhD in Education, worked at the time of the preparation of this study as a Head of Degree Programme (professional special needs teacher & student counsellor education) in The School of Professional Teacher Education, Häme University of Applied Sciences, Finland.

Jussi Silvonen

Jussi Silvonen, PhD and docent (adjunct professor) in Psychology, is an emeritus senior researcher in the University of Eastern Finland and current editor-in-chief for the Finnish Journal of Psychology.