Measuring individual and work-related correlates of employees’ well-being with a short form of the psychological general well-being index: A pilot validation study in Slovenia

Abstract

BACKGROUND:

The improvement of psychic well-being and mental health prevention in the workplace are high on the European agenda.

OBJECTIVE:

The study aimed to perform a construct validation of the previously developed short forms of the Psychological General Well-Being Index (PGWBI) on the working population and to identify the possible individual and work-related characteristics that co-shape individual perceptions of well-being.

METHODS:

A cross-sectional, quantitative study design was utilized. A standardized instrument and its four short forms developed in previous studies were used. The sample included 376 employees from larger companies of one Slovenian region. Exploratory and confirmatory factor analysis and multiple linear regression were performed.

RESULTS:

The short form that shows the best fit in this study sample is the PGWBI-B with RMSEA = 0.063; CFI = 0.986; SRMR = 0.026; NNFI = 0.977 and CI = 0.982. The significant independent variables were educational level (p < 0.001) and the average number of working hours per week (p = 0.001).

CONCLUSIONS:

The study proved that working conditions are correlated with employees’ well-being. Further validation of the PGWBI short forms should be performed with parallel and work context-specific measures of well-being to assess their concurrent validity. The monitoring of general employee well-being should become a regular practice in work organisations.

Keywords

Mental health psychological strain workplace construct validation

Dr Simona Hvalič-Touzery (PhD in Social work/Msc and BSc in Sociology) is an Assistant Professor at the Faculty of Social Sciences, University of Ljubljana. Prior to her current position, she was employed at the Angela Boškin Faculty of Health Care (ABFHC) and the Anton Trstenjak Institute. At ABFHC she worked as a Vice-Dean for Research and development. She lectured and performed research in the field of sociology of health and disease, health promotion and gerontology. Her specialisms are social gerontology and gerontechnology. Currently she works as a researcher and coordinator of projects focusing on smart solutions for active ageing.

Assist. Prof. Dr. Katarina Babnik has a PhD in Psychology and started her career in practice. She worked in the field of HRM for ten years. After obtaining her PhD, she was employed at the University of Primorska, where she lectured and performed research in the field of personnel development, occupational health, organizational psychology (occupational culture, occupational identity, socialization), and older workers. Currently she is employed at the University of Ljubljana, Faculty of Arts, Department of Psychology, where she lectures courses in the field of work and organizational psychology.

Sanela Pivač (M.Sc. (Nur), B.Sc. (Nur), Senior lecturer) is the acting dean at the Angela Boškin Faculty of Health Care (ABFHC). She completed specialist studies in General Management and obtained a Master’s in Nursing. She is a doctoral candidate at the Faculty of Medicine, University of Ljubljana. Her Public Health PhD thesis is in its final stage. She lectures on the first level of the study program in Nursing. She is a mentor and co-mentor to students when they prepare their theses at the first and second level of the study program in Nursing. She is active as a researcher and a project coordinator.

Dr Brigita Skela-Savič (PhD/MSc/BSc/RN) is Full Professor in Research, EBP and Management in health care area. She was a dean of faculty and international consultant for nursing education/development. She leads PhD and Master Programs at ABFHC, and is a guest professor on doctoral study programs at Slovenian universities. She is a member of the Board of Directors at ICN (2017 –2021), leader of Nursing Now in Slovenia (2019 - 2020) and member of editorial boards of scientific journals. She has reached the title Research Councilor with more than 70 scientific articles in scientific journals and participated in more than 60 research and development projects.

1 Introduction

Frequent and rapid changes in economic environments and organisational changes affect working conditions, leading to an increase in work intensity, work pressure and changes in employees’ perceptions of work security [1, 2], and their mental health [3]. The Third European Company Survey (ECS) in 2013 found that workplace well-being and establishment performance are associated; establishments with high workplace well-being are likely to have high establishment performance [4]. Moreover, researchers and research reviews [5] confirm that “poor working conditions, commonly referred to as ‘job stressors,’ can increase the risk for developing both clinical and sub-clinical disorders, including depression, anxiety, burnout, and distress”. It is therefore not surprising that, according to the European Foundation for the Improvement of Living and Working Conditions (Eurofound) and the European Agency for Safety and Health at Work (EU-OSHA) [6] survey, nearly 80% of managers in the European Union (EU) express concern for the work-related stress of employees and their well-being. The 2013 European Labour Force Survey found that among workers in the EU-27 and in Slovenia, 28.1% and 32.2%, respectively, were exposed to factors that can adversely affect mental well-being [7]. Several European documents have emphasized this issue. For example, Leka et al. [8] analysed 94 obligatory and non-binding policies with the conclusion that there is not enough awareness about mental health in the workplace and the importance of the preventive action. Monitoring the level of well-being of the working population is one of the first steps in health prevention in the workplace. The aim of our study was to validate one of the short well-being instruments on a sample of the Slovenian workforce and to determine the individual and work-related characteristics that might play an important role in employees’ well-being perceptions.

1.1 Theoretical background

1.1.1 The concept of well-being and its measures

Well-being is a broad concept and, over the past decades, important progress has been made in its description and explanation. In everyday conversation, the term has two general meanings: objective (material) and subjective (psychic), with the first describing individuals’ objective living conditions, such as income, housing arrangements and other indicators of welfare [9] and the second describing their experience and evaluations of life in general [9, 10]. Subjective perceptions of well-being are at the centre of attention of psychological science; beneath this scientific interest is the goal to explain and predict mental health [11]. General (psychic) well-being is described with two different, although correlated, constructs [12, 13]: subjective [14] and psychological well-being [15]. Subjective well-being is defined as individuals’ “overall evaluations of their lives and their emotional experiences” [14] and describes the level of happiness and how good someone feels [16]. The concept of psychological well-being is based less on the indicators and the ‘degree’ of well-being and more on the question of what constitutes “essential features of well-being” or what it means to be a “self-actualized, individuated, fully functioning or optimally developed” [17] person. Besides subjective and psychological well-being, two additional constructs describe a good life and mental health, namely the quality of life and wellness [11].

The diversity of constructs that define positive functioning results in the diversity of instruments assessing well-being. The systematic review performed by Cook et al. [11] identified 42 structured, self-rated well-being instruments, with substantial differences in “length, psychometric properties, and their conceptualization and operationalization of well-being.” For example, subjective well-being has three facets that are measured with separate instruments assessing life satisfaction, positive affect, and negative affect [14], although some studies assess only the specific component of well-being and some the whole construct [18]. Ryff [19] developed the Scales of Psychological Well-Being, which consist of six subscales (autonomy, environmental mastery, purpose in life, personal growth, positive relationships with others, and self-acceptance), with studies utilising one or more subscales [13].

The Psychological General Well-Being Index (the PGWBI) developed by Dupuy and colleagues [20] is a structured self-report scale. Operationalisation of the scale is based on an understanding of well-being as an individual’s life evaluation incorporating positive (optimism, enjoyment, self-control) and negative characteristics (depression, anxiety) that, together, reveal the level of that individual’s quality of life [10]. The PGWBI consists of 22 items, distributed across six subscales [20]: Anxiety (5 items), Depressed Mood (3 items), Positive Well-Being (4 items), Self-Control (3 items), General Health (3 items), Vitality (4 items), and a General Well-Being Index (all 22 items). The PGWBI therefore does not assess the psychological well-being construct as defined and operationalised by Ryff [15, 19], but rather mental health as the presence of positive and absence or low levels of negative self-evaluations and states, and some previous works [21 –24] describe the PGWBI as a quality-of-life measure. The PGWBI has been abundantly applied as the measure of the dependent or target variable in health-related studies [21 , 25–27]. In addition to being a measure of “health-related quality of life” and an “outcome indicator in clinical trial” [2], the PGWBI has been, albeit less often, applied to healthy population samples, such as students [28, 29], the working population [30], and older adults [31]. The reliability and validity of the first and the revised versions of the instrument and their subscales have been assessed internationally [1, 2], with studies showing their criterion (concurrent) validity with a high correlation of the PGWBI with other measures and indicators of mental health [24, 28]. Adequate sensitivity of the instrument to detect differences in well-being between patients with various health-related problems and the general population has also been demonstrated [21]. Study findings have also revealed that the PGWBI is a sensitive measure of changes in the level of well-being as a result of targeted interventions [27, 32]. Some studies [24 , 33] have questioned the validity of the six subscales of the PGWBI. The questionable validity of the dimensional structure of the PGWBI and the length of the scale stimulated researchers to develop and test its short forms [23 , 33]. To our knowledge, previous validation studies of the PGWBI short forms have not been performed on the general working population. Therefore, the first goal of this study was to validate different short forms of the PGWBI on a sample of the working population.

1.1.2 Well-being and the work context

Several theories and models try to explain the relation between work, stress, and well-being [34]. Most of them are based on the assumption that an individual’s health is the result of the fit between a person and his or her work environment [35]. The concept of working conditions covers different characteristics of the work environment, such as working time, remuneration, physical conditions, and mental demands in the workplace [36]. Working conditions that cause distress and lead to deterioration of mental and physical health represent psychological risk factors [6] or psychological hazards [8] that have a negative impact on employees’ mental health, mainly through the experience of work-related stress [8]. Psychological risk factors can be classified into two major categories (2): (i) job characteristics and the nature of work, and (ii) social and organisational context. Psychosocial risk factors include high job strain, poor effort-reward balance, high job insecurity, long working hours, low control over work, monotonous work, fixed-term employment contracts, high emotional demands, poor work-life balance, time pressure, and violence at work [6]. The third category of factors affecting work-related stress and well-being are individual characteristics of employees. Dispositional and acquired individual characteristics have a direct and an indirect (moderating and mediating) influence on the perceived work-related stress and well-being [37]. Some of the demographic variables with a possible relation to employees’ well-being are: gender [38], age [38], educational level [38 –40], and income [38].

In the light of the challenges faced by the labour market, employees’ well-being has to become an important concern. The goals of our study were to: (i) perform a construct validation of the short forms of the PGWBI on the working population and to (ii) identify possible individual and work-related characteristics that co-shape individual perceptions of well-being.

2 Methods

2.1 Design

A cross-sectional, non-experimental quantitative explorative research design was employed.

2.2 Instrument

As a measure of well-being, a structured self-report instrument, the Psychological General Well-Being Index (the PGWBI) [1 , 41], was used, since it “targets people’s self-representations of an aspect of their general well-being” [20], which was at the forefront of our study. In addition, it was selected because it has been widely used for several decades and validated by numerous studies [20]. The Mapi Research Trust gave the permission to use it and adapt it in the Slovenian language. The PGWBI consists of 22 items, measuring six dimensions: Anxiety (5 items), Depressed Mood (3 items), Positive Well-Being (4 items), Self-Control (3 items), General Health (3 items), Vitality (4 items), and a General Well-Being Index (all 22 items). The original 6-point scale (range: 0–5) by item was used, making the maximum score 110 with the direction of the score being the same for all. 0 represents the answer that reflects poor well-being, while 5 represents the answer that reflects a good well-being. A high total score indicates good quality of life while a low score indicates poor quality of life [1]. Previous exploratory dimensional analysis [24, 28] utilised the principal component method with varimax rotation that extracted different three-factor solutions [33]. A principal component analysis with varimax rotation and exploratory factor analysis (EFA) with promax method was performed in our study on 22 items of the GPWBI. The Kaiser-Meyer-Olkin (KMO) Measure of Sampling Adequacy and Bartlett’s test of sphericity statistics (Bartlett’s) were adequate (KMO = 0.96; Bartlett’s statistics p < 0.000). EFA yielded three factors with Eigen values greater than one. The three-factor solution explained 57.77% of variance, with the first factor explaining the greater proportion of variance (51.09%). The three factors were highly correlated; the lowest correlation, 0.65, was found to exist between the first and second factors and the highest, 0.71, between the second and third factors. High correlation between factors might imply a higher order unidimensional structure of the PGWBI scale, as has been implied in previous studies [33] and confirmed by Lundgren-Nilssonet et al. [32] and Testa et al. [33]. Lundgren-Nilssonet et al. [32] performed a Rasch analysis and a factor analysis on 22 items of the PGWBI scale to conclude that the scale measured only one dominant construct of well-being when the items from the six underlying domains were treated as six ‘super’ items. Since several short forms of PGWBI have been developed [23 , 33], one of the goals of our study were to perform a construct validation on a sample of working population.

We tested four different short forms of the PGWBI: the PGWB-S [23], the PGWBI-A, the PGWBI-B [33], and the QGBEP-R [29] with EFA and confirmatory factor analysis (CFA) (LISREL 10.20) [42]. Each of the short forms consists of six items; one from each of the subscales or dimensions of the original PGWBI scale (Table 1).

Table 1
Presentation of the PGWBI short forms

Short form Items from original PGWBI Subscales in the original PGWBI

PGWB-S Item 05 ANX

Item 06 V

Item 07 DM

Item 18 SC

Item 20 PWB

Item 21 V

PGWBI-A Item 3 DM

Item 4 SC

Item 6 V

Item 8 ANX

Item 10 GH

Item 20 PWB

PGWBI-B Item 9 PWB

Item 11 DM

Item 13 GH

Item 18 SC

Item 19 ANX

Item 21 V

QGBEP-R Item 1 PWB

Item 2 GH

Item 3 DM

Item 4 SC

Item 5 ANX

Item 6 VT

Short form	Items from original PGWBI	Subscales in the original PGWBI
PGWB-S	Item 05	ANX
	Item 06	V
	Item 07	DM
	Item 18	SC
	Item 20	PWB
	Item 21	V
PGWBI-A	Item 3	DM
	Item 4	SC
	Item 6	V
	Item 8	ANX
	Item 10	GH
	Item 20	PWB
PGWBI-B	Item 9	PWB
	Item 11	DM
	Item 13	GH
	Item 18	SC
	Item 19	ANX
	Item 21	V
QGBEP-R	Item 1	PWB
	Item 2	GH
	Item 3	DM
	Item 4	SC
	Item 5	ANX
	Item 6	VT

Note. Short forms are presented according to the year of publication. Legend: ANX, Anxiety; DM, Depressed Mood; PWB, Positive Well-Being; SC, Self-Control; GH, General Health; V, Vitality; PGWB-S developed by Grossi et al., [23], PGWBI-A, PGWBI-B developed by Testa et al., [33], QGBEP-R developed by Pereira et al. [29].

Demographic and work-related variables. For research purposes, 11 items were utilised to establish individual demographic characteristics and work-related characteristics. Additionally collected data included information on gender, age, educational level, type of employment (contract for an indefinite period vs. fixed-term contract), length of employment (overall), self-reported number of average working hours per week, work position, and self-reported number of sick leave days.

2.3 Participants

Eleven large companies (with more than 200 employees) of one Slovenian region were invited to participate in the research; of these, five companies responded. A total of 2,588 questionnaires were administered (census) and 456 were returned. In further analysis, we excluded the questionnaires that were not completely filled in, thus there is no missing data. The final sample size was 376. In terms of gender representation, 194 (51.6%) respondents were female. On average, respondents were 39.6 years old (SD = 9.4). A total of 16.8% of respondents held a managerial position, 66.2% worked in manufacturing (not in a managerial position), and 17.0% worked in administration (not in a managerial position). The average length of employment was 17.5 years (SD = 10.6). The majority of respondents had an employment contract for an indefinite period (85.1%). A total of 34.8% of respondents had taken sick leave over the previous year (Table 2). According to educational background, 30% of employee respondents had tertiary education. On average, respondents reported working 42.37 hours per week (SD = 4.39), with a minimum of 20 hours and a maximum of 60 hours.

Table 2
Respondents’ individual and work-related demographic information

n (%) M SD Min-Max

Total number of respondents 376 (100)

Gender: Male/Female 182 (48.4) / 194 (51.6)

Total Mean Age/Age range 39.6 9.4 19–63

Level of education

Primary education or lower 39 (10.4)

Secondary education–vocational school 95 (25.3)

Secondary education–senior or technical secondary school 129 (34.3)

Tertiary education–diploma or bachelor’s degree 102 (27.1)

Tertiary education–master’s degree or PhD 11 (2.9)

Employment contract

Contract for an indefinite period 320 (85.1)

Fixed-term contract 56 (14.9)

Work position

Managerial position at the company level 4 (1.1)

Managerial position in the working unit 35 (9.3)

I manage and supervise the work of executive workers 24 (6.4)

I am employed in administration and have no managerial position 64 (17.0)

I am employed in manufacturing and have no managerial position 249 (66.2)

Average length of employment (in years) 44.9 10.8 26–65

Average number of working hours per week 42.4 4.4 20–60

Average number of sick leave days in the last year among employees who took sick leave (n = 131) 11.6 23.9 1–180

	n (%)	M	SD	Min-Max
Total number of respondents	376 (100)
Gender: Male/Female	182 (48.4) / 194 (51.6)
Total Mean Age/Age range		39.6	9.4	19–63
Level of education
Primary education or lower	39 (10.4)
Secondary education–vocational school	95 (25.3)
Secondary education–senior or technical secondary school	129 (34.3)
Tertiary education–diploma or bachelor’s degree	102 (27.1)
Tertiary education–master’s degree or PhD	11 (2.9)
Employment contract
Contract for an indefinite period	320 (85.1)
Fixed-term contract	56 (14.9)
Work position
Managerial position at the company level	4 (1.1)
Managerial position in the working unit	35 (9.3)
I manage and supervise the work of executive workers	24 (6.4)
I am employed in administration and have no managerial position	64 (17.0)
I am employed in manufacturing and have no managerial position	249 (66.2)
Average length of employment (in years)		44.9	10.8	26–65
Average number of working hours per week		42.4	4.4	20–60
Average number of sick leave days in the last year among employees who took sick leave (n = 131)		11.6	23.9	1–180

Legend: n, Number of answers; SD, Standard deviation; %, Share; M, Mean value; Min, Minimum; Max, Maximum.

2.4 Data collection

The relevant company boards adopted a decision on each company’s participation. Respondents received information about different aspects of the study; their rights on voluntary participation and withdrawal from the study at any time as well as an explanation of their privacy and confidentiality rights.

2.5 Data processing

Data analysis was performed in two phases. The previously developed short PGWBI forms were first analysed with EFA (maximum likelihood, varimax rotation). A one-dimensional structure of each of the four proposed short forms of GPWBI was tested with CFA (maximum likelihood). Sun [43] suggests the following evaluation criteria of CFA models in construct validity evaluation studies: SRMR (Standardized Root Mean Square Residual), NNFI (Non-Normed Fit Index), Mc or CI (McDonald’s Centrality Index; Centrality Index), RMSEA (Root Mean Square Error of Approximation), and CFI (Comparative Fit Index). LISREL output does not give the information on CI; therefore, we computed the CI statistics with the formula presented in Jaccard and Wan [44]:

CI = exp(-1/2d)

where d is:

d = (χ²_model –df_model)/ N_sample

The model’s χ² and df (df _shortforms = 9) were gathered for each of the four tested models (four models for each of the short form of PGWBI) from the LISREL output.

The evaluation of the goodness of fit of each model was assessed through the following recommended values of fit statistics [44, 45]: χ² p > 0.05; RMSEA < 0.07; CFI > 0.95; SRMR < 0.05; NNFI > 0.95; CI > 0.90.

The short version of the PGWBI that yielded the most appropriate fit statistics was utilised in further analysis. We used descriptive analysis and multiple linear regression to assess the correlations between individual and work-related variables and the score on the chosen short form of the PGWBI. Before performing a further analysis for assessing the differences between individual and work-related characteristics, two categorical variables (Level of education and Work position) were replaced with dichotomous variables, since the frequencies of specific categories were low (e.g. “Managerial position” for the Work position variable, “Tertiary education–master’s degree or PhD” for the Level of education variable) or had a low discriminant value (e.g. “Managerial position in the working unit” and “I manage and supervise the work of executive workers.”). The two new dichotomous variables were: Level of education, with two categories: (i) vocational or lower: grouped categories “Primary school” and “Secondary education–vocational school” (N = 134) and (ii) secondary or higher: grouped categories “Secondary education–senior or technical secondary school”, “Tertiary education–diploma or bachelor’s degree” and “Tertiary education–master’s degree or PhD” (N = 242); and Work position, with two categories: (i) managerial position: grouped categories “Managerial position at the company level”, “Managerial position in the working unit”, “I manage and supervise the work of executive workers” (N = 63) and (ii) non-managerial position: “I am employed in administration and have no managerial position” and “I am employed in manufacturing and have no managerial position” (N = 313).

Data were analysed using statistical software SPSS 26.0 and LISREL 10.20. Statistical significance was set at the p < 0.05 level.

3 Results

3.1 Construct validation of the PGWBI short forms

The EFA performed on each of the short forms of the PGWBI resulted in one factor solution for each form with adequate KMO measure and Bartlett’ statistics (for all four forms KMO > 0.8 and Bartlett’s statistics p < 0.00) that explains 58.75% (the PGWB-S), 47.57% (the PGWBI-A), 52.64% (the PGWBI-B), and 46.01% (the QGBEP-R) of variance. In the next step, we analysed the goodness of fit measures of one factor solution of each short form with the CFA. Table 3 presents the goodness of fit measures of unidimensional models of the PGWBI short versions. The short form that shows the best fit in this study sample is the PGWBI-B [33] with RMSEA = 0.063; CFI = 0.986; SRMR = 0.026; NNFI = 0.977 and CI = 0.982, but with significant p-value of the model χ² (χ² = 22.390; p = 0.0077).

Table 3
Evaluation of unidimensional short forms of the PGWBI

Short Form Goodness of Fit Measures

χ²(df = 9) RMSEA CFI SRMR NNFI CI

(p) (90% confidence interval)

PGWB-S 57.610 0.120 0.966 0.028 0.944 0.937

(Grossi et al., 2006) (p = 0.0000) (0.092; 0.151)

PGWBI-A 32.195 0.083 0.970 0.032 0.951 0.969

(Testa et al., 2016) (p = 0.0002) (0.053; 0.115)

PGWBI-B 22.390 0.063 0.986 0.026 0.977 0.982

(Testa et al., 2016) (p = 0.0077) (0.030; 0.096)

QGBEP-R 83.499 0.149 0.926 0.044 0.877 0.905

(Pereira et al., 2018) (p = 0.0000) (0.121; 0.179)

Short Form		Goodness of Fit Measures
PGWB-S	57.610	0.120	0.966	0.028	0.944	0.937
(Grossi et al., 2006)	(p = 0.0000)	(0.092; 0.151)
PGWBI-A	32.195	0.083	0.970	0.032	0.951	0.969
(Testa et al., 2016)	(p = 0.0002)	(0.053; 0.115)
PGWBI-B	22.390	0.063	0.986	0.026	0.977	0.982
(Testa et al., 2016)	(p = 0.0077)	(0.030; 0.096)
QGBEP-R	83.499	0.149	0.926	0.044	0.877	0.905
(Pereira et al., 2018)	(p = 0.0000)	(0.121; 0.179)

Note. Short forms are presented according to the year of publication. Legend: χ2, Model Chi-square; RMSEA, Root Mean Square Error of Approximation; CFI, Comparative Fit Index; SRMR, Standardized Root Mean Square Residual; NNFI, Non-Normed Fit Index; CI, McDonald’s Centrality Index.

Table 4 presents the completely standardised lambda (factor loadings) and R² (variance extracted) of the PGWBI-B. Table 4 shows that standardized factor loadings range between 0.550 (item 13) and 0.890 (item 19). For a construct convergent validity, we followed Hair et al.’s [46] rule of thumb: “individual standardized factor loadings should be at least 0.5, and preferably 0.7”. In further psychometric evaluations of the PGWBI-B, the possibilities of replacing item 13 with another relevant one should be considered. The reliability (Cronbach’s alpha) of the short form PGWBI-B was 0.860.

Table 4

Standardized factor loadings and variance extracted for the PGWBI-B items

PGWBI-B item	Standardizedlambda (λ)	R²
Item 9	0.680	0.462
Item 11	0.689	0.475
Item 13	0.550	0.303
Item 18	0.718	0.516
Item 19	0.890	0.792
Item 21	0.782	0.612
AVE		52.60%

Legend: R², R-Square, AVE, Average variance extracted.

For further analysis, the summed scale [46] PGWBI-B was computed for each respondent as a sum of the raw scores (ranging from 0 to 5) on each of the six items comprising the PGWBI-B scale, with the minimum possible score 0 and the maximum 30. This summed scale score represented the well-being score for each participant.

3.2 Employees’ well-being and the role of individual and work-related characteristics

Table 5 shows descriptive statistics of the 6-item version of the PGWBI-B. The total score of the PGWBI-B shows that, on average, respondents self-reported as being moderately well (M = 20.344; SD = 5.340). Item 21 (original item by Chassany et al. [20], “I felt tired, worn out, used up, or exhausted during the past month”), was evaluated lowest by the respondents with an average score of three (M = 3.080; SD = 1.219), which, on the response scale anchors, means that respondents evaluated the frequency of such states as “some of the time”. Item 11 received the highest average evaluation of four (M = 4.069; SD = 1.157), showing that, on average, respondents evaluated the item “Have you felt so sad, discouraged, hopeless, or had so many problems that you wondered if anything was worthwhile during the past month?” (original item by Chassany et al. [20]) with the response anchor “A little bit.”

Table 5
Descriptive statistics of the six items and the summed PGWBI-B score

Item Min Max M SD

Item 9 0 5 3.223 0.973

Item 11 0 5 4.069 1.157

Item 13 0 5 3.380 1.266

Item 18 0 5 3.412 1.197

Item 19 0 5 3.294 1.124

Item 21 0 5 3.080 1.219

PGWBI-B score 5 30 20.441 5.340

Item	Min	Max	M	SD
Item 9	0	5	3.223	0.973
Item 11	0	5	4.069	1.157
Item 13	0	5	3.380	1.266
Item 18	0	5	3.412	1.197
Item 19	0	5	3.294	1.124
Item 21	0	5	3.080	1.219
PGWBI-B score	5	30	20.441	5.340

Legend: Min, Minimum; Max, Maximum; M, Mean value; SD, Standard deviation.

The skewness and kurtosis statistics for the PGWBI-B score show that the overall score on the GPWBI-B scale is distributed close to normal, since the Skewness and Kurtosis measures were lower than +/–1.0 (Skewness = –0.334, SE = 0.126; Kurtosis = –0.537, SE = 0.251).

A t-test was used to establish the differences in the PGWBI-B score between Gender, Level of education, Work position, Type of employment (“Contract for an indefinite period” or “Fixed-term contract”) and Sick leave over the last 12 months (“yes” or “no”). Before computing the t-tests, Levene’s test of equality of variance was performed, which confirmed the equality of variances for all variables. T-test results did not reveal any significant differences on the PGWBI-B score between genders (t = –1.718, p = 0.087), respondents holding a managerial or non-managerial position (t = –0.124, p = 0,901), respondents that had been on sick leave over the previous 12 months and those that had not (t = –1.760, p = 0.079), or those that have an employment contract for an indefinite period or a fixed-term contract (t = 1.776, p = 0.077). The difference in the summed PGWBI-B score was significant at the p < 0.05 level only in regard to Level of education (t = –3,736; p = 0.000). Respondents with a completed vocational secondary school or lower (i.e. primary school) self-reported a lesser degree of well-being (lower PGWBI-B score) (M = 19.082, SD = 5.686) compared to those who had completed a senior or technical secondary school or higher (M = 21.194, SD = 4.993). Pearson’s correlation test indicated that the summed scores of the PGWBI-B did not correlate with age (r = –0.049, p = 0.344) and length of employment (in years) (r = –0.69, p = 0.062), but they did correlate with working hours per week (r = –0,158, p = 0.002).

Linear regression was used to identify the relationship between the summed PGWBI-B score and independent variables for which a significant correlation between the independent and dependent variable (GPWBI-B score) was established in the previous analysis. Results are shown in Table 6. The obtained significant independent variables were educational level (p < 0.001) and the average number of working hours per week (p = 0.001), which together explained 6% of the variance of the PGWBI-B score.

Table 6

Multiple linear regression model for the summed PGWBI-B score

	PGWBI-B score (R² =0.060)
Characteristics –Model	b	SE _b	β	p
Educational level	1.040	0.283	0.187	< 0.001
(dichotomous)
The average number	–0.208	0.062	–0.171	0.001
of working hours
per week (metric)

Legend: R², R-Square; b, Regression coefficient; SE_b, Standard regression coefficient error; β, Standard regression coefficient; p, p value.

4 Discussion

In Slovenia, mental and behavioural disorders are the third-ranked reasons for sick leave in the age group 45–65 years. The number of adults perceiving daily stress related to working conditions has been increasing over the years (14). In line with these findings, our research sought to validate a short instrument that could be applied by employers for regular monitoring of employees’ well-being. The CFA identified the PGWBI-B [33] short form of the PGWBI as having the most appropriate goodness of fit measure. Nevertheless, it should be noted that this was one of the first studies validating the construct validity of various short forms of the PGWBI on a sample of the working population. Further studies are needed and criterion validity evaluation should be performed. Well-being might be conceptualised as general or domain-specific [14]. In the context of workplace, domain-specific well-being is best represented by constructs such as job-related attitudes, stress [14], quality of work-life [47], and burnout [48]. In this study, we measured the general well-being of the working population without assessing the domain-specific well-being constructs, since previous studies revealed that the PGWBI instrument is adequate for studies and clinical interventions of stress exhaustion, burnout [32], and work-family-work conflict [30]. In their systematic review, Salvagioni et al. [49] showed the physical, psychological and occupational detrimental consequences of job-related burnout and the necessity of its early identification in workplaces as a step towards healthier working environments. Measuring the general well-being of the working population is therefore a necessary occupational health prevention activity, in addition to monitoring context-specific work- and job-related constructs, such as job satisfaction [50, 51].

The second goal of our study was to determine the individual and work-related characteristics associated with the level of well-being. The results show that the respondents evaluated their well-being as moderately good. Key findings show that the well-being score can be explained with respondents’ educational level and the average number of working hours per week. Workers with lower levels of education are more vulnerable in relation to mental health. Previous studies confirm these results. A study performed on a large sample of employees from Slovenia [39] found that psychophysical health complaints are less frequent among more highly educated people (feelings of depression, anxiety, social contact issues). Lunau et al. [40] analysed the associations between educational level and work stress in a sample of employees from 16 European countries and confirmed the association between lower education and higher levels of work stress in all countries. Such results can be partly explained by Karasek’s job demand and control model [6]. Employees with lower education have more monotonous, repetitive tasks and lower levels of job autonomy [6]. These working characteristics may have a negative impact on psychological well-being and stress, especially if employees are subjected to high job demands (quantity and quality of work) [52]. Working hours are one of the variables describing working conditions with an impact on employees’ mental health. Our study confirmed that workload correlates with the score on the PGWBI-B. Baka [53] found that employees’ working hours positively correlate with depression, job burnout and physical symptoms. Similarly, Eurofound [6] warns that long working hours are a major component of work-related stress. Linear regression explained only a small percentage of the variation. More context-specific well-being constructs, focused on work and the working environment, could be explained by the studied variables to a greater extent, but this is not always the aim in human resource management. Monitoring employees’ general well-being should be implemented as a regular human resource practice in the field of health promotion.

4.1 Limitations

There are some limitations to this study. A higher response rate would have been desired. Linear regression explained only a small percentage of the variation. However, our findings are still important since we proved that working conditions are a factor impacting employees’ well-being. Although the short form PGWBI-B [33] showed the most adequate values on the construct confirmation criteria, further validation should be performed with parallel and work context-specific measures of well-being to assess their concurrent validity. Although short forms of instruments might have several advantages for participants [54], the validation of presented short forms of PGWBI, should be based on additional validation studies of the original scale and its short versions.

5 Conclusion

Despite its limitations, our study confirms that educational level and long working hours have an important effect on employees’ well-being. The validated short form of the PGWBI might be used in organisational settings to monitor the level of well-being of employees. Well-being assessment should become a regular monitoring practice in the work context with the goal to raise awareness of the importance of mental health, to develop and implement prevention activities, and to support or assist those employees with already lower levels of well-being.

Footnotes

Acknowledgments

The research was conducted within the project “The establishment of preventive mental health promotion programs (ANIMA SANA) (public tender EEA Grants/Norway Grants 2009–2014)”. The project was funded by the Ministry of Economic Development and Technology of Slovenia and EEA Grants/Norway Grants and coordinated by the Upper Carniola Development Agency, Slovenia. Permission to conduct the research was obtained from the Senate Committee for Science, Research and Development at the Angela Boškin Faculty of Health Care on May 19th, 2015. The serial number of Approval is 3/9 in the academic year 2014/2015.

References

Dupuy

. The Psychological generalWell-Being (PGWB) Index. In: Wenger NK, Mattson ME, Furberg CD, Elinson J, editors. Assessment of quality of life clinical trials of cardiovascular therapies. S.l.: Le Jacq Publishing; 1984. pp. 170-83.

Revicki

, Leldy

, Howland

. Evaluating the psychometric characteristics of the Psychological General Well-Being Index with a new response scale. Qual Life Res. 1996;5(4):419–25.

ten Have

, van Dorsselaer

, de Graaf

. The association between type and number of adverse working conditions and mental health during a time of economic crisis (2010-2012). Soc Psychiatry Psychiatr Epidemiol. 2015;50(6):899–907.

Kankaraš

, van Houten

. Third European Company Survey –overview report: Workplace practices –patterns, performance and well-being. Luxembourg: European Foundation for the Improvement of Living and Working Conditions (Eurofound); 2015.

Memish

, Martin

, Bartlett

, Dawkins

, Sanderson

. Workplace mental health: An international review of guidelines. Prev Med. 2017;101:213–22.

Eurofound—European Foundation for the Improvement of Living and Working Conditions, EU-OSHA — European Agency for Safety and Health at Work. Psychosocial risks in Europe: Prevalence and strategies for prevention: A joint report from the European Foundation for the Improvement of Living andWorking Conditions and the European Agency for Safety and Health at Work. Luxembourg: Publications Office of the European Union; 2014.

European Commission, Eurostat. Labour force survey in the EU, candidate and EFTA countries: Main characteristics of national surveys. Luxembourg: Publications Office of the European Union; 2015.

Leka

, Jain

, Iavicoli

, Di Tecco

. An evaluation of the policy context on psychosocial risks and mental health in the workplace in the European Union: Achievements, challenges, and the future. BioMed Res Int. 2015;2015:1–18.

Lincoln

, Taylor

, Chae

, Chatters

. Demographic correlates of psychological well-being and distress among older African Americans and Caribbean black adults. Best Pract Ment Health. 2010;6(1):103–26.

10.

Wilkinson

. Work life balance and psychological wellbeing in men and women. Auburn: Auburn University; 2013.

11.

Cooke

, Melchert

, Connor

. Measuring well-being: A review of instruments. Couns Psychol. 2016;44:730–57.

12.

Musek

. Dimenzije psihičnega blagostanja. Antrophos. 2008;40(1/2):139–60.

13.

Babnik

, Benko

, von Humboldt

. Ryff’s Psychological Well-Being Scale. In: Gu D, Dupre ME, editors. Encyclopedia of gerontology and population aging. Cham: Springer; 2019.

14.

Diener

, Heintzelman

, Kushlev

, Tay

, Wirtz

, Lutes

, et al. Findings all psychologists should know from the new science on subjective well being. Canadian Psychology. 2017;58:87–104.

15.

Ryff

. Beautiful ideas and the scientific enterprise: Sources of intellectual vitality in research on eudaimonic well-being. In: Vittersø J, editor. Handbook of Eudaimonic Well-Being. Cham: Springer; 2016. pp. 95-108.

16.

Taneva

. What is psychological well-being and how it changes throughout the employment cycle? In: Monteiro I, Iguti AM, editors. Trabalho, saúde e sustentabilidade: Construindo a cidadania=Work, health and sustainability: Building citizenship. Campinas: UnicampBFCM; 2016. pp. 83-90.

17.

Ryff

. Psychological well-being revisited: Advances in the science and practice of eudaimonia. Psychother Psychosom. 2013;83(1):10–28.

18.

Bartels

. Genetics of wellbeing and its components satisfaction with life, happiness, and quality of life: A review and meta-analysis of heritability studies. Behav Genet. 2015;45(2):137–156.

19.

Ryff

. Happiness is everything, or is it? Explorations on the meaning of psychological well-being. J Pers Soc Psychol. 1989;57(6):1069–81.

20.

Chassany

, Dimenäs

, Dubois

, Wu

. The Psychological General Well-Being Index (PGWBI) User Manual. Lyon: MAPI Research Institute; 2004.

21.

Bianchi

, Marchesini

, Nicolino

, Graziani

, Sgarbi

, Loguercio

, et al. Psychological status and depression in patients with liver cirrhosis. Dig Liver Dis. 2005;37(8):593–600.

22.

Davison

, Bell

, LaChina

, Holden

, Davis

. The relationship between self-reported sexual satisfaction and general well-being in women. J Sex Med. 2009;6(10):2690–7.

23.

Grossi

, Groth

, Mosconi

, Cerutti

, Pace

, Compare

, et al. Development and validation of the short version of the Psychological General Well-Being Index (PGWB-S). Health Qual Life Outcomes. 2006;4:88.

24.

Wool

, Cerutti

, Marquis

, Hervié

, Cialdella

. Psychometric validation of two Italian quality of life questionnaires in menopausal women. Maturitas. 2000;35(2):129–142.

25.

Gartoulla

, Bell

, Worsley

, Davis

. Moderate-severely bothersome vasomotor symptoms are associated with lowered psychological general wellbeing in women at midlife. Maturitas. 2015;81(4):487–92.

26.

Lanciano

, Curci

. Does emotions communication ability affect psychological well-being? A study with the Mayer–Salovey–Caruso Emotional Intelligence Test (MSCEIT) v2.0. Health Commun. 2015;30(11):1112–21.

27.

Macri

, Limoni

. Artistic activities and psychological well-being perceived by patients with spinal cord injury. Arts Psychother. 2017;54:1–6.

28.

Gaston

, Vogl

. Psychometric properties of the General Well-Being Index. Qual Life Res. 2005;14(1):71–5.

29.

Pereira M da

, Antunes

, Barroso

, Correia

, Brito

, Monteiro

. Adaptation and validation of the psychological General Well-Being Index: Confirmatory factor analysis of the short version. J Nurs Referência Rev Enferm Referência. 2018;4:9–11.

30.

Matthews

, Kath

, Barnes-Farrell

. A short, valid, predictive measure of work-family conflict: Item selection and scale validation. J Occup Health Psychol. 2010;15(1):75–90.

31.

Narushima

, Liu

, Diestelkamp

. Lifelong learning in active ageing discourse: Its conserving effect on wellbeing, health and vulnerability. Ageing Soc. 2018;38(4):651–75.

32.

Lundgren-Nilsson

, Jonsdottir

, Ahlborg

, Tennant

. Construct validity of the psychological general well-being index (PGWBI) in a sample of patients undergoing treatment for stress-related exhaustion: A rasch analysis. Health Qual Life Outcomes. 2013;11:2.

33.

Testa

, Civilotti

, Di Fini

, Veglia

, Rossetto

, Boncinelli

. Development of two equivalent short forms of the Psychological General Well-Being Index: PGWBI-A and PGWBI-B. Test Psychom Methodol Appl Psychol. 2016;23:149–66.

34.

Marchand

, Juster

R-P

, Durand

, Lupien

. Work stress models and diurnal cortisol variations: The SALVEO study. J Occup Health Psychol. 2016;21:182–93.

35.

Mark

, Smith

. Stress models: A review and suggested new direction. In: Houdmont J, Leka S, editors. Occupational Health Psychology, European Perspectives On Research, Education and Practice. Vol. 3. Nottingham: Nottingham University Press; 2008. pp. 111-44.

36.

International Labour Organization –ILO [homepage on the Internet]. Working conditions. [cited 2019 Avg 7]. Available from: https://www.ilo.org/global/topics/workingconditions/lang–en/index.htm.

37.

De Jonge

, Dormann

. Whyismyjob so stressful? Characteristics, processes and models of stress at work. In: Chmiel N, Fraccaroli F, Sverke S, editors. An introduction to work and organizational psychology: An international perspective, 3rd ed. Hoboken: John Wiley & Sons; 2017.

38.

Cifre

, Vera

, Signani

. Women and men at work: Analyzing occupational stress and well-being from a gender perspective. Rev Puertorriqueña Psicol. 2015;26:172–91.

39.

Boštjančič

, Slana

. Psychophysical workload of workers in Slovenia. In: Petelin A, Šarabon N, Prosen M, Žvanut, editors. Health of the working age population: Proceedings 2017. Koper: University of Primorska Press; 2017. pp. 9-22.

40.

Lunau

, Siegrist

, Dragano

, Wahrendorf

. The association between education and work stress: Does the policy context matter? PLoS One. 2015;10(3):1–17.

41.

Grossi

, Compare

. Psychological general well-being index (PGWB). In: Michalos AC, editor. Encyclopedia of quality of life and well-being research. Dordrecht: Springer Netherlands; 2014, pp. 5152-6.

42.

Jöreskog

, Olsson

, Wallentin

. Multivariate Analysis with LISREL. Basel: Springer; 2016.

43.

Sun

. Assessing goodness of fit in confirmatory factor analysis. Meas Eval Couns Dev. 2005;37:240–56.

44.

Jaccard

James

, Wan

. LISREL approaches to interaction effects in multiple regression. Thousand Oaks: Sage Publications; 1996.

45.

Hooper

, Coughlan

, Mullen

. Structural equation modelling: Guidelines for determining model fit. Electron J Bus Res Methods. 2008;6(1):53–60.

46.

Hair

, Black

, Babin

, Anderson

, Tatham

. Multivariate Data Analysis. 7th ed. Upper Saddle River: Pearson Prentice Hall; 2009.

47.

Sirgy

, Lee

. Work-Life Balance: A Quality-of-Life Model. Appl Res Qual Life. 2016;11(4):1059–82.

48.

Maslach

, Leiter

. Burnout. In: Fink G, editor. Stress: Concepts, cognition, emotion and behaviour. San Diego: Elsevier Academic Press; 2016. pp. 351-7.

49.

Salvagioni

DAJ

, Melanda

, Mesas

, González

, Gabani

, Andrade

SM de

. Physical, psychological and occupational consequences of job burnout: A systematic review of prospective studies. PloS One. 2017;12:e0185781.

50.

Supriyanto

. Compensation effects on job satisfaction and performance. Hum Syst Manag. 2018;37(3):281–5.

51.

Boštjančič

, Petrovčič

. Exploring the relationship between job satisfaction, work engagement and career satisfaction: The study from public university. Hum Syst Manag. 2019;38(4):411–22.

52.

Chung-Yan

. The nonlinear effects of job complexity and autonomy on job satisfaction, turnover, and psychological well-being. J Occup Health Psychol. 2010;15(3):237–51.

53.

Baka

. Does job burnout mediate negative effects of job demands on mental and physical health in a group of teachers? Testing the energetic process of Job Demands-Resources model. Int J Occup Med Environ Health. 2015;28(2):335–46.

54.

Levant

, Hall

, Weigold

, McCurdy

. Construct distinctiveness and variance composition of multi-dimensional instruments: Three short-form masculinity measures. Journal of Counseling Psychology. 2015;62(3):488–502.

Short Form		Goodness of Fit Measures
	χ²(df = 9)	RMSEA	CFI	SRMR	NNFI	CI
	(p)	(90% confidence interval)
PGWB-S	57.610	0.120	0.966	0.028	0.944	0.937
(Grossi et al., 2006)	(p = 0.0000)	(0.092; 0.151)
PGWBI-A	32.195	0.083	0.970	0.032	0.951	0.969
(Testa et al., 2016)	(p = 0.0002)	(0.053; 0.115)
PGWBI-B	22.390	0.063	0.986	0.026	0.977	0.982
(Testa et al., 2016)	(p = 0.0077)	(0.030; 0.096)
QGBEP-R	83.499	0.149	0.926	0.044	0.877	0.905
(Pereira et al., 2018)	(p = 0.0000)	(0.121; 0.179)

Measuring individual and work-related correlates of employees’ well-being with a short form of the psychological general well-being index: A pilot validation study in Slovenia

Abstract

BACKGROUND:

OBJECTIVE:

METHODS:

RESULTS:

CONCLUSIONS:

Keywords

1 Introduction

1.1 Theoretical background

1.1.1 The concept of well-being and its measures

1.1.2 Well-being and the work context

2 Methods

2.1 Design

2.2 Instrument

2.5 Data processing

3 Results

3.1 Construct validation of the PGWBI short forms

Table 5 Descriptive statistics of the six items and the summed PGWBI-B score Item Min Max M SD Item 9 0 5 3.223 0.973 Item 11 0 5 4.069 1.157 Item 13 0 5 3.380 1.266 Item 18 0 5 3.412 1.197 Item 19 0 5 3.294 1.124 Item 21 0 5 3.080 1.219 PGWBI-B score 5 30 20.441 5.340

4.1 Limitations

5 Conclusion

Footnotes

Acknowledgments

References

Table 5
Descriptive statistics of the six items and the summed PGWBI-B score

Item Min Max M SD

Item 9 0 5 3.223 0.973

Item 11 0 5 4.069 1.157

Item 13 0 5 3.380 1.266

Item 18 0 5 3.412 1.197

Item 19 0 5 3.294 1.124

Item 21 0 5 3.080 1.219

PGWBI-B score 5 30 20.441 5.340