Emotional intelligence has become a key factor in well-being, in highlighting the importance of intellectual functioning and information processing (Salovey & Mayer, 1990). The definition of emotional intelligence in research is well documented as the ability to recognise one’s own and others’ emotions, to differentiate between feelings and use emotional information which guides thinking, behaviour and performance (Boyatzis, 2009; Goleman, 1995; Joseph & Newman, 2010; Mayer & Salovey, 1997; Salovey & Mayer, 1990).

The present study investigates the role of emotional intelligence and exhaustion within the job demands-resource model (JD-R), regarded as a heuristic model based on the premise of two sets of working conditions that may produce health impairment and motivation (Adil & Baig, 2018). The health hindrance (distress) is known as the reduced energy process, and this represents the imbalance between high job demands and low job resources and the motivational process indicates that low job resources prevent the ability to cope, whilst high job demands lead to reduced accomplishments (Adil & Baig, 2018). Emotional intelligence abilities function as important psychological resources, such as perceiving and managing emotions can compensate for excessive demands (Adil & Baig, 2018). Psychological resources are depleted when the emotional nature of work in call centres does have direct outcomes on exhaustion and personal accomplishment (Harry, 2015).

Research proves that perceiving, managing, monitoring and utilising own and others’ emotions are deemed positive in people coping better in life when in control of their emotions (Zerach & Levin, 2018). In order for individuals to regulate emotions, the effectiveness of perceiving, utilising and managing emotions favours the intellectual and emotional development of individuals. Emotional intelligence contributes to subjects understanding and performing their actions with greater efficiency (Hanafi, 2016; Salovey & Mayer, 1990). Individuals can effectively deal with emotions from within as well as from others and in turn manage this emotional information by becoming self-aware and having the ability to perceive emotions. Self-management results are the ability to manage emotions and stimuli from the environment (Hanafi, 2016; Salovey & Mayer, 1990).

Research also indicates that emotional intensity is positively related to emotional exhaustion and depersonalisation, suggesting that occupational burnout is mainly caused by stressful interaction with clients (Teinlin et al., 2017). The concept of positive affectivity has received much attention from various researchers all suggesting that positive affectivity is negatively related to emotional exhaustion. Emotional intelligence should be approached from an ability as well as trait models (Zerach & Levin, 2018). Ability models capture employees’ ability to use, and the ability of what a person is capable of doing. The construct of trait emotional intelligence focuses on a constellation of emotions, namely the extent to which employees identify, understand and regulate one’s own and others’ emotions (Hanafi, 2016; Salovey & Mayer, 1990). Results in research reveal that trait emotional intelligence especially within call centres in Africa is useful in identifying individual differences and more likely to appraise stressful situations and be able to cope in such situations (Harry, 2015).

Within the JD-R model, specific work characteristics/demands such as call centres do associate with burnout/exhaustion that necessitates emotional labour, which is characterised by the constant pressure to meet demands, resulting in psychological strain (Adil & Baig, 2018). These demands require sustained psychological effort that is associated with high work pressure. The emotional nature of work in call centres does have direct outcomes on burnout dimensions of exhaustion and personal accomplishment, as it affects individual traits such as emotional intelligence and affective states of emotions experienced at work (Harry, 2015). In research conducted in call centres what has been established is negative affectivity was significantly associated with higher levels of emotional exhaustion (Szczygiel & Bazinska, 2013). Emotional exhaustion is caused by work pressure and a loss of energy. Empirical research indicates burnout to be caused by many organisational variables, amongst which it is important to highlight the client-related stressor especially given the changes that are taking place in call centre in Lagos, Nigeria, and Johannesburg and Durban, South Africa (Schäfer, Becker, King, Horsch, & Michael, 2019).

Research into the third dimension of burnout shows that the professional efficacy points to a lack of personal fulfilment, which results in employees leaving the organisation, a characteristic of many call centres in Africa (Anwar & Graham, 2019). On the other hand, having a sense of efficacy does include aspects that a worker would consider as being meaningful and worthwhile: accomplishments and abilities can serve as a buffer to emotional exhaustion, as previously mentioned (Anwar & Graham, 2019).

Researchers have reviewed the social perspective of exhaustion in which people utilise the interpersonal relation by responding and perceiving others through the process of motivation and emotions, especially when coping with emotional arousal. The importance of managing others’ emotions, perceiving others’ emotions and understanding emotions of customers are important internal resources for dealing with exhaustion and depersonalisation within call centres in Africa (Schäfer et al., 2019).

The sense of meaningfulness, regarded as a motivational component is important as a moderating mechanism that explains the use of emotional intelligence resources such as managing, utilising and perceiving emotions in call centre work (Richardson & Ratner, 2006; Rohani, Abedi, Sundberg, & Langius-Eklöf, 2015). This is based on the conditional interaction with work-related outcomes, constituting the extent of belief that things in life are worthwhile and there is a good reason for any individual to care about what happens and wishes to cope with all indicators of health and well-being (Richardson & Ratner, 2006; Rohani et al., 2015). Authors have reported in the previous work on the moderating effects of the sense of meaningfulness by analysing the degree of moderation with different variables such as subjective perceptions of psychological health and wellbeing and experience with stress symptoms (Richardson & Ratner, 2006; Rohani et al., 2015).

Research in Africa regarding the moderating effects of the sense of meaningfulness and exhaustion and depersonalisation has been established as having a negative effect. This indicates that the sense of meaningfulness as a protective function modifies the negative effects of stress (Johnston et al., 2013). In other research conducted, sense of meaningfulness had negative associations with exhaustion and depersonalisation. This implies individuals will start to assess the availability of resources to increase feeling of control (Eriksson & Mittelmark, 2017). The sense of coherence can be better in predicting perceptions and managing emotions, based on areas of presumed overlapping such as perceptions of external and emotional stimuli, and making sense of these stimuli. Several authors have indicated that the role of perceiving and managing emotions and the sense of meaningfulness has links to wellbeing (Burnett, 2019; Shkoler & Tziner, 2017; Szczygiel & Bazinska, 2013). A further exploration of the sense of meaningfulness (sense of coherence) can be regarded as another psychological resource that can spark employees to feel engaged in their work and to find a sense of meaning in remaining in the company (Adil & Baig, 2018).

This study investigates the interaction effects between call centre agents’ emotional intelligence (i.e. managing, utilising and perceiving emotions) and their sense of meaningfulness (as a moderating mechanism) in predicting their exhaustion and professional efficacy.

The profile of the participants consisted of 409 permanently employed call centre employees in three of the largest outsourced financial call centres in Lagos, Nigeria, and Johannesburg and Durban in South Africa. The participants consisted of 66% women and 34% men.

Three assessment instruments were used for this study. The Assessing Emotions Scale (AES), developed by Schutte et al. (1998), was used to measure the following four emotional intelligence traits: perceptions of emotion (10 items; e.g. ‘I am aware of my emotions’), managing own emotions (9 items; e.g. ‘I have control over my emotions’), managing others’ emotions ( 8 items; e.g. ‘I like to share my emotions with others’) and utilisation of emotions (6 items; e.g. ‘When my mood changes, I see new possibilities’). Validity studies on the AES have justified the various underlying constructs of the four subscales (Ciarrochi, Chan, & Caputi, 2000). In terms of reliability (internal consistency), Ciarrochi et al. (2000) reported Cronbach’s α coefficient to be 0.55 (moderate) to 0.78 (high).

The Orientation to Life Questionnaire (OLQ) was developed by Antonovsky (1993) to measure meaningfulness (8 items; e.g. ‘You anticipate that your personal life in the future will be totally without meaning or purpose versus full of meaning and purpose’). Reliability and validity studies have been empirically tested across different cultures and confirm Cronbach’s α ranging from 0.51 to 0.90. (Eriksson & Mitelmark, 2017).

The Maslach Burnout Inventory developed by Maslach, Jackson and Leiter (1996) was used to measure the two dimensions of burnout, exhaustion (5 items; e.g. ‘Working all day is really a strain for me’) and professional efficacy (6 items; e.g. ‘I have accomplished many worthwhile things in this job’). Reliability studies have shown the MBI to have Cronbach’s α ranging from 0.70 to 0.80 (Edwards et al., 1997).

The study has been carried out strictly in accordance with the Policy for Research Ethics of the University of South Africa (UNISA) and ethical clearance was obtained from the University Research Ethics committee. The author took great care that the research was conducted with the highest integrity, taking into account UNISA’s Policy for Infringement and Plagiarism. The human resource manager of the call centres in Africa granted permission to conduct the research study. Each questionnaire included a covering letter inviting the respondents to participate voluntarily in the study and assuring them that their individual responses would remain anonymous and confidential. After 500 questionnaires were sent out, the research procedure yielded 409 useable questionnaires (response rate = 82%).

Moderated regression analysis, applying the Hayes PROCESS procedure (Hayes, 2018), was used to achieve the research aim. The more stringent, bias-corrected bootstrapping technique (Hayes, 2018) with the 95% lower level (LLCI) and upper level (ULCI) range not including zero, was used to examine the significance of the interaction effects. Cohen’s f² was used to assess the practical effect of significant interaction effects.

Table 1 presents the means, standard deviations, internal reliability coefficients of the variables and bi-variate correlations.

As shown in Table 1, the internal consistency reliability for the scales ranged between (> 0.61 and > 0.87). The bivariate correlations between meaningfulness and utilising emotions, perceptions of emotions and managing emotions variables ranged between r ≥ 0.18 and r ≤ 0.79 (p ≤ 0.01; small practical effect) suggesting that multicollinearity was not a threat to the findings (r < 0.80). The exhaustion variable negatively correlated with meaningfulness, utilising emotions, perceptions of emotions and managing emotions (r ≥ −0.12 and r ≤ −0.38).

Professional efficacy variable correlated positively with meaningfulness, utilising emotions, perceptions of emotions and managing emotions (r ≥ 0.12 and r ≤ 0.30).

Table 2 presents moderated regression analysis. Four models with a more stringent bootstrapping approach (Hayes, 2018) were used. Model 1, comprising of three variables: managing others’ emotions moderated by sense of meaningfulness in predicting exhaustion; Model 2, comprising of three variables: utilising emotions moderated by sense of meaningfulness in predicting exhaustion; Model 3, comprising of three variables: perceptions of emotions moderated by sense of meaningfulness in predicting exhaustion; Model 4, comprising of three variables: managing own emotions moderated by sense of meaningfulness in predicting exhaustion.

According to Table 2, only two regression models were significant: Model 1, F_p = 28.198, p = 0.001 and Model 3, F_p = 27.542, p = 0.001. The adjusted R² ranged between 0.16 and 0.17 (moderate practical effect).

Significant interaction effects were only observed in terms of Model 1 (managing others’ emotions x meaningfulness: β = −0.40; LLCI = −0.63; ULCI = −0.16; p = 0.01; f² = 10.58 (small practical effect) and Model 3 (perceptions of emotions x meaningfulness: β = −0.45; LLCI = −0.72; ULCI = −0.19; p = 0.001; f² = 11.32 (small practical effect). Figures 1 and 2 illustrate the interaction effects.

Figure 1 displays the interaction effect, that is managing others’ emotions x sense of meaningfulness in predicting exhaustion, shows that the participants high on managing others’ emotions and high sense of meaningfulness experienced significantly lower levels of exhaustion than those who had low levels of managing others’ emotions and sense of meaningfulness. Those participants who had a significantly low sense of meaningfulness also had a significantly high level of exhaustion irrespective of their high level of managing others’ emotions.

Figure 2 displays the interaction effect, that is perceptions of emotions x sense of meaningfulness in predicting exhaustion, shows that the participants high on perceptions of emotions and high sense of meaningfulness experienced significantly lower levels of exhaustion than those who had low levels of perceptions of emotions and sense of meaningfulness. Those participants who had a significantly low sense of meaningfulness also had a significantly high level of exhaustion irrespective of their high level of perceptions of emotions.

Table 3 presents moderated regression analysis. Four models with a more stringent bootstrapping approach (Hayes, 2018) were used. Model 1, comprising of three variables: managing others’ emotions moderated by sense of meaningfulness in predicting professional efficacy; Model 2, comprising of three variables: utilising emotions moderated by sense of meaningfulness in predicting professional efficacy; Model 3, comprising of three variables: perceptions of emotions moderated by sense of meaningfulness in predicting professional efficacy; Model 4, comprising of three variables: managing own emotions moderated by sense of meaningfulness in predicting professional efficacy.

According to Table 3, only two regression models were significant: Model 1: F_p = 24.59, p = 0.035 and Model 3: F_p = 17.078, p = 0.013. The adjusted R² ranged between 0.11 and 0.17 (moderate practical effect). Significant interaction effects were only observed in terms of Model 1 (managing others’ emotions x meaningfulness: β = −0.16; LLCI = −0.30; ULCI = −0.01; p = 0.04; f² = 4.47 (small practical effect) and Model 3 (perceptions of emotions x meaningfulness: β = −0.21; LLCI = −0.37; ULCI = −0.04; p = 0.01; f² = 6.25, (small practical effect). Figures 3 and 4 illustrate the interaction effects.

Figure 3 displays the interaction effect, that is managing others’ emotions x sense of meaningfulness in predicting professional efficacy, shows that the participants low on managing others’ emotions and with a low sense of meaningfulness experienced significantly lower levels of professional efficacy than those who had high levels of managing others’ emotions and a high sense of meaningfulness. Those participants who had high levels of managing others’ emotions and a significantly high sense of meaningfulness also experienced higher levels of professional efficacy than those with a lower sense of meaningfulness.

Figure 4 displays the interaction effect, that is perceptions of emotions x sense of meaningfulness in predicting professional efficacy, shows that the participants low on perceptions of emotions and with a low sense of meaningfulness experienced significantly lower levels of professional efficacy than those who had high levels of managing others’ emotions and a high sense of meaningfulness. Those participants who had high levels of perceptions of emotions and a significantly high sense of meaningfulness also experienced higher levels of professional efficacy than those with a lower sense of meaningfulness.

Interventions to develop sense of meaningfulness, managing others’ emotions and perceptions of emotions may help to reduce levels of exhaustion and increase professional efficacy in the call centre industry. The research suggests that sense of meaningfulness, managing others’ emotions and perceptions of emotions as an important wellbeing variable when improving the wellness of call centre agents in Africa.

This study could not generalise the findings to other occupational contexts as it was confined to the call centre work environment only with a sample of predominantly of single, employed black females. The self-reporting approach was followed and the association between the variables could also have been influenced by common method variance. Though, self-reports were probably the most accurate means as they involved internal psychological processes.