Fisher (1993) describes boredom as a fleeting unpleasant state, arising from inter alia under-stimulation, where the employee experiences a lack of interest and finds it difficult to concentrate on the present activity or situation. The essential elements of the boredom phenomenology include the individual’s inability or difficulty maintaining focus and concentration (Fisher, 1993; Loukidou et al. 2009). In addition, it is associated with an unpleasant feeling, which causes the individual to search for variety in order to alleviate this negative state (Hamilton, Haier & Buchsbaum, 1984; Loukidou et al. 2009). Workplace boredom is similar to boredom but refers to the experience of boredom as described above, contextualised within the domain of work.

Thus, workplace boredom can be defined as a task- or activity-related, unpleasant, dissatisfying and de-activating emotion due to low arousal or the under-stimulating nature of the job (Fisher, 1993; Mikulas & Vodanovich, 1993). Workplace boredom is a temporary state of mind where one may feel bored one moment but not the next (Fisher, 1993). However, boredom at work becomes problematic if it is experienced on a frequent basis. It differs from other affective states (e.g. work engagement) as bored employees often feel unchallenged, thus making the situation or the job seem increasingly meaningless (Van Tilburg & Igou, 2012). Meaning encompasses sense of purpose and values that provide a sense of justification and one’s sense of self-worth (Baumeister, 1991).

Some researchers on workplace boredom have attributed the cause of workplace boredom to the work situation itself (Hill & Perkins, 1985). Traditionally, it was assumed that job characteristics such as monotony or the repetitive nature of tasks inevitably led to workplace boredom (Hill & Perkins, 1985). However, according to Shackleton (1981), not everyone working within repetitive jobs is likely to experience workplace boredom. Research suggests that employees’ appraisal of the (work) situation determines their response in terms of boredom (Lazarus, 1991; Weiss & Cropanzano, 1996), for example, the employee’s capability versus task complexity mentioned earlier. Seen from this perspective, the job characteristics and work environment are considered to be potential sources of workplace boredom (Loukidou et al. 2009). These job characteristics refer specifically to task variety, autonomy and utilisation of skills (Fisher, 1993; Loukidou et al. 2009; Smith, 1981).

Interest in the boredom domain has led to the development of various scales in order to measure the phenomenon. However, only a single scale (the DUBS) was available that focused specifically on the employee’s experience of boredom in the work environment. Earlier scales, such as the Job Boredom Scale (JBS) (Grubb, 1975; Lee, 1986), focused primarily on the antecedents of workplace boredom with reference to an under-stimulating work environment and excluded the affective, cognitive and behavioural manifestations of boredom itself (Vodanovich, Wallace & Kass, 2005). Alternatively, dispositional boredom scales (Farmer & Sundberg, 1986) (i.e. susceptibility to boredom as a personality trait such as sensation-seeking) took various boredom responses into account but did not relate these responses to the work environment. Reijseger et al. (2012) combined the strengths of both the JBS and dispositional scales to create a new scale known as the DUBS. The DUBS measures work-related boredom and focuses on the employee’s affective, cognitive and behavioural responses with regard to an under-stimulating work environment.

Reijseger et al. (2012) developed the DUBS in the Netherlands. The process of developing the DUBS included construct conceptualisation, item adaption (combining of two existing boredom scales), psychometric evaluation and further refinement. The items of the DUBS were primarily aimed at measuring the experience and manifestation of workplace boredom (cf. Baker, 1992; Hill & Perkins, 1985). An eight-item scale was designed at first. These DUBS items were adapted from the Boredom Proneness Scale (Farmer & Sundberg, 1986) and the JBS (Grubb, 1975; Lee, 1986). The DUBS measures feelings, thoughts and behaviours (or the absence of behaviours) that are prevalent when boredom is experienced at work. These include the perception of time passing slowly and feeling bored (Hill & Perkins, 1985), task-unrelated thoughts (Damrad-Frye & Laird, 1989) and inclinations to engage in non-work-related activities (Baker, 1992).

Reijseger et al. (2012) used three convenience samples that included respondents from various organisational sectors. All item scores correlated significantly; hence, a unidimensional construct was apparent. Furthermore, confirmatory factor analyses (CFAs) provided the necessary evidence to confirm a one-factor structure consisting of six items. Two items were dropped from the original scale because they appeared to be ambiguous and unsound. The resulting six-item scale showed an acceptable reliability as the Cronbach’s alpha indicated a high level of internal consistency (α = 0.80) (Reijseger et al. 2012). Thus, based on previous work on the DUBS in the Netherlands, two hypotheses were formulated:

Hypothesis 1: Workplace boredom, as assessed with the DUBS, is a one-dimensional construct.

Hypothesis 2: The workplace boredom scale is reliable (i.e. Cronbach’s α ≥ 0.70).

The validity of a measuring instrument is considered important as this influences the precision, interpretation and usefulness of the findings (Foxcroft & Roodt, 2009; Westen & Rosenthal, 2003). When determining the validity of an instrument, one method is to investigate its convergent validity. Convergent validity refers to the extent to which two theoretical constructs that are expected to be related to one another are in fact related (Campbell & Fiske, 1959). It is expected that significant correlations between the DUBS and other constructs such as work engagement (negative correlation), organisational commitment (negative correlation) and work underload (positive correlation) will be evident.

Work underload and workplace boredom: In the present study, work underload referred to the employee’s perception of work in terms of reversed role overload (Beehr, Walsh & Taber, 1976). In this context, work underload can be described as having too little work to do within a given timeframe (Larson, 2004) or having work that does not live up to the individual’s competence, skills and knowledge (Fineman & Payne, 1981; Spector & Jex, 1998).

Due to having too little to do, or experiencing one’s job as too simple and unchallenging, workplace boredom is experienced. Workplace boredom may therefore occur when quantitative ‘underload’ is experienced; that is, when tasks are considered to be monotonous, unchallenging or repetitive (Fisher, 1993). Thus, the expectation exists that there is a positive relationship between workplace boredom and work underload, which leads to the following:

Hypothesis 3: Workplace boredom and work underload are positively related.

Work engagement and workplace boredom: The link between workplace boredom and work engagement suggests that boredom at work impairs employee productivity and well-being (Whiteoak, 2014). Work engagement is typically described as ‘a positive, fulfilling work-related state of mind that is characterised by vigour, dedication, and absorption’ (Schaufeli, Salanova, Gonzalez-Romá & Bakker, 2002, p. 74). Contrarily, bored employees experience a dissatisfying negative state, which is usually associated with negative outcomes and withdrawal from work (Bruursema, Kessler & Spector, 2011; Game, 2007). It appears that the presence of workplace boredom decreases work engagement and vice versa (Reijseger et al. 2012). Warr and Inceoglu (2012) support the notion that work engagement is negatively related to workplace boredom, and they describe workplace boredom (an unpleasant state of passiveness) as the polar opposite of work engagement (a pleasant and activated state), in line with Warr’s model of affective well-being (Warr, 1990). Similarly, Schaufeli and Salanova (2014) hypothesised and showed that work engagement is inversely related to workplace boredom. Hence, the following hypothesis is formulated:

Hypothesis 4: Workplace boredom has a negative relationship with work engagement.

Organisational commitment and workplace boredom: Although organisational commitment has affective, continuance and normative components (Allen & Meyer, 1990; Meyer, Allen & Smith, 1993), the focus in the current study was on the affective component. The reason is that it appears to be the most consistent commitment component that is associated with other organisational variables (cf. Meyer, Stanley, Herscovitch & Topolnytsky, 2002). According to Allen and Meyer (1990), the affective component taps the employee’s identification and emotional connection with the organisation as well as participation in the organisation.

When workplace boredom is prevalent, employees experience their job as dissatisfying; therefore, they are less committed to the organisation and willing to leave their jobs (Kass et al. 2001; Reijseger et al. 2012). Reijseger et al. (2012) found a negative relationship between workplace boredom and organisational commitment. This finding indicates that employees who experience workplace boredom also experience less organisational commitment – indicating a negative exchange relationship.

Predictive validity is an alternative form of validity that can be useful in a validation study. It refers to the precision with which a measure can predict future behaviour, responses or category status (Foxcroft & Roodt, 2009), that is, the predictive regression based on theoretical argument. Predictive validity is typically associated with a longitudinal research design, where cause and effect can be more definitively demonstrated. However, cross-sectional data allow the researcher to begin investigations of the regressions with regard to a one-directional relationship (De Vos, Strydom, Fouché & Delport, 2012; O’Dwyer & Bernauer, 2014). For the purpose of this study, predictive validity was considered by means of a mediation model – see Figure 1 for a representation of this model. As argued above, work underload may result in an employee feeling under-challenged, leading to boredom and in turn to a possible decrease in work engagement and organisational commitment. Based on this reasoning, the final two hypotheses were formulated.

To date, no research on workplace boredom has been conducted within the South African context. South Africa is a unique multicultural nation where different cultural backgrounds, ethnicities and values are found (Foxcroft & Roodt, 2009). The historic context of this nation and the various adversities faced (e.g. the after-effects of apartheid) might affect the manner in which a psychological phenomenon is perceived (Rodrik, 2008). As a result, the values, norms and assumptions of the South African workforce could differ from those in other countries and could thus influence the experience of a phenomenon such as workplace boredom (Foxcroft & Roodt, 2009). The suitability of utilising a measuring instrument from another country could be considered questionable, if the measuring instrument was not validated in the specific country, presenting the possibility of questionable results (Foxcroft & Roodt, 2009). Furthermore, the Employment Equity Act (Government Gazette, 1998) prohibits the use of psychological assessment without appropriate evidence of reliability and validity within this context. Therefore, the purpose of this research was to investigate the psychometric properties of a workplace boredom measure for use within South Africa, as the first step to assist future researchers on the topic in this context.

Permission and consent to gather data were obtained from logistic and manufacturing organisations, as well as from the individuals who participated in the survey. Hard copies of the surveys were printed and handed out to participating employees by the researcher. Each survey consisted of a cover letter explaining the purpose of the research. Emphasis was placed on anonymity, confidentiality and voluntary participation with no negative repercussions.

The study sample consisted of a combination of participants from logistics and manufacturing organisations (N = 490). The mean age of participants was 39.16 (standard deviation = 11.35). About half (51%) of the sample consisted of female participants. Participants who were married or living with a partner constituted 58% of the sample. Most of the respondents possessed a university degree (34%), followed by a Grade 12 qualification (22%). The number of participants who spoke Western Germanic languages was 327 (68%), and 156 spoke an African language (31%).

All of the items in the survey were presented in English only. The Cronbach’s reliability coefficient for each scale in this section was based on previous studies with the scales. The Cronbach’s alpha coefficients for the current study are presented in Table 3.

Workplace boredom was measured with the DUBS, developed by Reijseger et al. (2012). This one-dimensional scale for measuring workplace boredom uses a five-point rating scale ranging from 1 (‘never’) to 5 (‘always’). Workplace boredom is measured by six items (e.g. ‘I feel bored at my job’ and ‘I tend to do other things during my work’). The scale has a reported Cronbach’s alpha coefficient of 0.80 (Reijseger et al. 2012).

Work engagement was measured in terms of items from the Utrecht Work Engagement Scale (UWES) (Schaufeli et al. 2006). This instrument typically consists of 17 items, which are all scored on a seven-point frequency-rating scale ranging from 0 (‘never’) to 6 (‘always’). Only items from the core components of work engagement (i.e. vigour and dedication items) were used as indicators of work engagement (Langelaan, Bakker, Van Doornen & Schaufeli, 2006). For the purposes of the current study, 11 items were used measuring vigour (e.g. ‘I can continue working for very long periods at a time’) and dedication (e.g. ‘I find the work that I do full of meaning and purpose’). The UWES has been used in South Africa with good reliability (Storm & Rothmann 2003).

Organisational commitment was measured using a five-point scale ranging from 1 (‘strongly disagree’) to 5 (‘strongly agree’). This scale is based on the items of the affective organisational commitment component by Allen and Meyer (1990) and five items were used (e.g. ‘I would be very happy spending the rest of my career with this organisation’). Cronbach’s alpha coefficients of above 0.70 have been reported for this scale (Allen & Meyer, 1990).

Job satisfaction was measured with the scale developed by Hellgren, Sjöberg and Sverke (1997). This three-item instrument uses a five-point scale ranging from 1 (‘strongly disagree’) to 5 (‘strongly agree’) to measure the individual’s satisfaction with his or her job (e.g. ‘I enjoy being at my job’). In the study by Hellgren et al. (1997), this scale reported a Cronbach’s alpha coefficient of 0.86, and in South Africa an alpha of 0.80 has been shown (Pienaar, Sieberhagen & Mostert, 2007).

Work underload a workload scale was used (Price, 2001; adapted from Kim, Price, Mueller & Watson, 1996). One reversed item for workload was used as the first item in the factor to reverse the polarity of the factor to work underload in the CFA (e.g. ‘My workload is not heavy on my job’). A five-point rating scale was used, which ranged from 1 (‘strongly disagree’) to 5 (‘strongly agree’).

Preliminary item analysis was performed in Mplus with the TYPE=BASIC function. This provided means, standard deviations, skewness, kurtosis and minimum and maximum values. Skewness and kurtosis would be considered problematic for any values below –2.00 or above 2.00 (George & Mallery, 2010). To investigate the reliability of the constructs, Cronbach’s alpha coefficients were calculated. Acceptable values for the reliability coefficients would have a lower-bound cut-off value of 0.70 (Bland & Altman, 1997). Moreover, structural equation modelling (SEM) methods were applied with Mplus 7.31 (Muthén & Muthén, 2014). The estimation method implemented was the maximum likelihood (ML) estimator for all analyses. Firstly, a CFA was implemented in order to investigate the factor loadings of the individual items of the latent workplace boredom variable. Secondly, a total measurement model was constituted to include all the study variables to constitute a structural model, based on latent variables, to investigate the correlations and relationships proposed in the study hypotheses. No correlation of error terms between observed indicators were forced or any parcelling strategies used. To evaluate the fit of all the models to the data, the following fit indices were considered (Cudeck & Browne, 1993; Van de Schoot, Lugtig & Hox, 2012):

The comparative fit index (CFI) (values of 0.90 and above).

Discriminant validity is established when evidence can be provided that the constructs of interest in a study can be shown to not be captured by other measures in the model (Hair, Black, Babin & Anderson, 2010). Thus, study variables should not correlate too highly and an undue influence in variation should not be evident between variables, which can distort results (e.g. r ≥ 0.85; Brown, 2015). In this study, the method described by Fornell and Larcker (1981) is used to establish discriminant validity. This method computes the average variance extracted (AVE) and the shared variance between the constructs involved. To pass this test, the AVE of the latent constructs should be higher than the shared variance between those constructs. Therefore, workplace boredom should be sufficiently correlated with the constructs as described above (H₃–H₅), but overlap should not be excessive and discriminant validity should exist.

The practical significance of correlation coefficients were set at 0.30 and above for a medium effect and at 0.50 and above for a large effect (Cohen, 1988). Overall statistical significance was considered at the 95% level (p ≤ 0.05).

To investigate the mediation model, the MODEL INDIRECT function in Mplus was used with 5000 bootstrap replications within the structural model specifications. The focus was therefore on the significance of the indirect effect and its 95% confidence interval (Rucker, Preacher, Tormala & Petty, 2011). Furthermore, the classification of the mediation model was done according to the guidelines of Zhao, Lynch and Chen (2010). Specifically, these guidelines indicate that the mediation model can be classified as an indirect-only mediation if the relationship from the independent variable to the mediator is significant and the relationship from the mediator to the outcome is also significant (synonymous to full mediation). However, if the relationship from the independent variable is also significant, the model is classified as a complementary mediation model (synonymous to partial mediation).