Demands–abilities fit (Edwards, 1996; Kristof-Brown et al., 2005) refers to the extent to which job requirements match the skills and abilities of the employee. The Theory of Work Adjustment (Bretz & Judge, 1994) emphasises that individuals will contribute longer in their jobs when there is D–A fit, and the job environment facilitates the use of their skills and abilities. Therefore, organisations make every effort to hire and retain employees with high D–A fit. Likewise, employees also strive for fit between their own perceived abilities, job demands and job resources (Greguras & Diefendorff, 2009). Although fit researchers agree that perceived D–A fit is associated with positive work outcomes, some fit studies have failed to find empirical support for this relationship (Astakhova, 2016; Oh et al., 2014).

Demands–abilities fit as a dimension of person–environment fit contributes to a belief that the working environment is conducive to what the organisation wants (Greguras & Diefendorff, 2009). Work roles that are aligned with individuals’ abilities and self-concepts should be associated with more meaningful work experiences (May et al., 2004). Fulfilling roles that are congruent with an individual’s strengths (Peterson & Seligman, 2004) contributes to the experience of meaningful work and engagement (May et al., 2004).

Beliefs about the function of work in life can shape the meaning of one’s work (Wrzesniewski & Tosti, 2005). Meaning in work is also described as the level of general significance that the experience of working has in the life of people at a given time (Bellah et al., 1985). The subjective experience of working is classified into three broad categories, namely, work as a job, work as a career and work as a calling (Bellah et al., 1985; Schwartz, 1994; Wrzesniewski et al., 1997). Employees who view their work as a job are only interested in the material benefits from work. They see their work as a means to acquire the resources needed to enjoy their time away from the job rather than an end in itself. These job holders do not express their significant interests and ambitions through their work (Parry, 2006; Wrzesniewski et al., 1997; Wrzesniewski, Dutton, & Debede, 2003).

Individuals who view their work as a career have invested in their work and mark their achievements not only through financial gains but also advancement in their careers (Parry, 2006; Wrzesniewski et al., 1997). In this case, meaning is derived from a perceived higher social standing and self-esteem, as well as increased power within the scope of one’s occupation (Bellah et al., 1985). Individuals who view their work as a career are happier than those who see their work as a job. However, they are less happy than those who regard their work as a calling (Dik & Duffy, 2008; Peterson et al., 2009).

Individuals with a calling orientation regard their work as inseparable from their life. In this case, work is not merely for financial gain or career advancement, but instead for the fulfilment that is possible by doing it (Peterson et al., 2009; Wrzesniewski et al., 1997, 2003). Work that employees feel called to do is usually seen as socially valuable – an end in itself – involving activities that may, but need not, be pleasurable and financially worthwhile (Bellah et al., 1985). Viewing work as a calling has benefits for the individual, the group and the organisation, including energy, life satisfaction and organisational commitment (Cameron et al., 2003; Peterson et al., 2009; Wrzesniewski et al., 2003).

Individuals have some control over the extent to which they experience their work as meaningful (Kahn & Heaphy, 2014). They might craft their jobs (i.e. to view their work as a calling) by changing the quality and amount of interaction with others or by seeing themselves as helpers of vulnerable others (including organisms). In this way, job crafters seek out audiences who can help them sustain desirable identities. Meaningful work results from the relations with those who benefit from their work and those who confirm its importance (Kahn & Heaphy, 2014).

The three ways in which people view their work are still largely unexplored in individuals who have a job in nature conservation. A qualitative study by Bunderson and Thompson (2009) showed that people who work with animals work for passion rather than for pay or advancement. They found that a sense of calling was grounded in a perceived connection between personal passion and capabilities and domains of work for which they are well-suited. Forsyth (1994) and Palmer and Bryant (1985) found a high level of job satisfaction among game wardens in America. One of the factors mentioned in this regard was the match between the outdoor orientation of wardens and the nature of their work (Palmer & Bryant, 1985). If the work that people do allows them to express themselves in work roles that are socially valuable, they will develop a heightened sense of meaningfulness and personal engagement. A sense of calling may offer the ‘strongest’ (Bellah et al., 1985, p. 66) path to meaningful work.

Current employees of protected areas in South Africa, including nature reserves, national parks and privately owned reserves, as well as people in nature-related jobs, such as training facilities for nature-based careers, were included as participants in the study. These employees consisted of management (including heads of departments), conservationists, educationists, researchers, tour guides and field staff. Data were gathered from these participant groups (N = 300) using a non-probability convenience sampling method (Sarantakos, 2013). A final number of 161 people completed the survey, resulting in a response rate of 53.67%. Table 1 describes the participants’ characteristics.

Table 1 shows that male participants comprised 42.90% and females 57.10% of the sample. Participants’ ages ranged from 19 to 75 (Mean = 38.89; SD = 12.92). Most participants (87.58%) had completed a qualification higher than matric, with 37.89% having a master’s degree or higher qualification. The length of service in the current job position varied from 1 to more than 20 years, whereas most participants (42.86%) had more than 10 years in a nature-related job. Most participants (64.60%) were permanently employed and 86.96% were South African citizens. English and Afrikaans were the home languages of 38.51% and 34.78% participants, respectively, whereas 26.72% participants spoke one of the African languages at home.

Data were collected using an online questionnaire. The first section of the questionnaire obtained demographic information about participants’ age, gender, language, years working in the current position in a nature-related industry, level of education, type of current employment and citizenship. In the second section of the questionnaire, standardised surveys were used to measure participants’ D–A fit, their purpose and meaning in work, work beliefs and work engagement.

The Demands–Abilities Fit Scale (DAFS; Greguras & Diefendorff, 2009) was used to measure the extent to which job requirements matched the skills and abilities of the employee. Three items developed by Cable and DeRue (2002) from the Person–Environment Fit Scales (Greguras & Diefendorff, 2009) were used. An example item is ‘The match is very good between the demands of my job and my personal skills’. The items required the respondent to answer on a scale ranging from 1 (strongly disagree) to 5 (strongly agree). Greguras and Diefendorff (2009) reported the reliability of this scale as α = 0.82.

The work beliefs of participants were measured by the Work–Life Questionnaire (WLQ) (Wrzesniewski et al., 1997). The WLQ is a self-report measure that classifies an individual’s work orientation into three main categories, namely, work as a job, career or calling (Wrzesniewski et al., 1997). The WLQ is divided into two parts. The first part consists of three paragraphs representing the three meanings of work. The respondent has to rate his or her level of association with each paragraph on a scale of 1 (very much like me) to 4 (not at all like me). The second part consists of 18 items formulated to substantiate the respondent’s answers on Part 1 of the questionnaire (e.g. ‘My primary reason for working is financial’). The items are rated on a Likert scale ranging from 1 (very much like me) to 4 (not at all like me). In previous research, Wrzesniewski et al. (1997) found the reliability of this instrument to be adequate. Van Zyl, Deacon and Rothmann (2010) reported Cronbach’s alpha coefficients to be ranging between 0.80 and 0.87 for the WLQ in a South African study.

The Work and Meaning Inventory (WAMI) (Steger et al., 2012) was administered to measure experiences of meaningful work. According to Steger et al. (2012), meaningful work consists of three dimensions, namely, sensing that work is a key avenue for making meaning, experiencing positive meaning in work and perceiving one’s work to serve some greater good. The WAMI consists of 10 items measuring three subscales, namely, meaning making through work (three items, e.g., ‘I view my work as contributing to my personal growth’), positive meaning (four items, e.g., ‘I understand how my work contributes to my life’s meaning’) and greater good motivations (three items, e.g., ‘The work I do serves a greater purpose’). The items are rated on a five-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree). Studies showed that meaningful work scores correlate with work-related and general well-being indices (Steger et al., 2012). Reliabilities ranging from 0.82 to 0.89 were obtained for the subscales.

Work engagement was measured by an adapted version of the Work Engagement Scale (WES) (May et al., 2004). The WES has nine items. For all items, a Likert scale ranging from 0 (never) to 6 (always) was used. The WES assessed the three dimensions of Kahn’s (1990) conceptualisation of work engagement, namely, cognitive engagement (e.g. ‘I am immersed in my work’), emotional engagement (e.g. ‘I am enthusiastic about my job’) and physical engagement (e.g. ‘At my work, I feel bursting with energy’). In a South African study, Olivier and Rothmann (2007) obtained an alpha coefficient of 0.72, which supports the reliability of the total scale.

The survey was constructed by an independent contractor for online surveys. The approved online survey was emailed to the participants from the convenience sample group. By including as many as possible employment groups and individuals in each group, the limitations that are usually inherent in convenience sampling (Wagner, Kawulich, & Garner, 2012) were likely to be adequately addressed. Participants completed the online survey and responses were electronically captured by the independent contractor, who forwarded a daily update of responses to the researcher. The data were prepared for statistical analyses with SPSS.

Data from the questionnaires were captured in SPSS 22 (IBM Corp., 2013). Following this, the dataset was screened for errors and outliers following the procedure outlined in Field (2015). Once this process was completed, descriptive statistics were calculated for all items, scales and subscales.

Given the relatively limited sample size, the structure of the four measuring instruments was investigated using exploratory factor analyses. Principal component analyses were used to estimate the number of components in each measuring instrument by considering the eigenvalues, percentage of variance explained and the scree plots (Field, 2015). Responses to each questionnaire were subjected to principal factor analysis with a direct oblimin rotation. According to Tabachnick and Fidell (2007), the following guidelines can be used to evaluate factor loadings: factor loadings higher than 0.71 are considered excellent, 0.63 very good, 0.55 good, 0.45 fair and 0.32 poor. A cut-off point of 0.40 was set for cross-loadings (Hair, Babin, Money, & Samouel, 2010, p. 364).

Cronbach’s alpha coefficients were computed to study the reliability of the measuring instruments. Descriptive statistics were computed to describe the data. Pearson’s correlation coefficients were used to specify the relationship between the variables. The practical significance of findings was assessed through effect sizes (Steyn, 2000). The practical significance of correlation coefficients was studied using the guidelines of Cohen (1988).

The discriminant validity of the measures was assessed in this study following a procedure suggested by Farrell (2010). A subscale or scale will have discriminant validity if it accounts for more variance in the observed variables associated with it than other variables in a model. The validity of indicators and the construct is questionable if this is not the case (Fornell & Larcker, 1981). The average variance explained (AVE) for each construct was compared with the shared variance between the constructs. Discriminant validity is supported if the AVE for a construct is greater than its shared variance with any other construct.

Furthermore, regression analyses were used in this study. First, standard multiple regression analyses were used to investigate the main effects of D–A fit and work beliefs on meaningful work. Second, standard multiple regression analyses were used to investigate the main effects of D–A fit and work beliefs on meaningful work and employee engagement. The following guidelines of Cohen (1988) were used to assess the practical significance of the explained variance: R² = 0.09 (medium effect) and R² = 0.25 (large effect).

Relative weight analysis (Tonidandel & LeBreton, 2015) was used to provide a partitioning of the variance among correlated predictors of meaningful work and engagement. PROCESS (version 2; Hayes, 2013) was used to assess indirect effects of antecedents of meaningful work on work engagement (via meaningful work). Using confidence intervals (CIs), this macro for SPSS estimates the indirect effects of X on Y through one or more mediator variable(s). Bias-corrected CIs (95% CI with 10 000 resamples) were used to assess whether indirect effects were different from zero (Hayes, 2013).

A principal component analysis was carried out on the three items of the DAFS. One factor (eigenvalue = 2.14) explaining 71.92% of the variance was extracted. The component loading and communalities (h²) were as follows: Item 1 = 0.76 (h² = 0.57), Item 2 = 0.91 (h² = 0.83) and Item 3 = 0.85 (h² = 0.74). This indicates that the three items of the DAFS are valid indicators of the construct.

A principal component analysis was carried out on the 21 items of the WLQ. Five factors with eigenvalues larger than one were extracted. The eigenvalues of the five factors and percentages of variance extracted were as follows: Factor 1 = 5.56 (26.49%); Factor 2 = 2.19 (10.44%), Factor 3 = 1.80 (8.55%), Factor 4 = 1.50 (7.15%) and Factor 5 = 1.21 (5.74%). Given that a three-factor structure was expected for the WLQ, it was decided to specify three factors. Next, a principal factor analysis with a direct oblimin rotation specifying three factors was carried out. Five items (WLQ7, WLC10, WLC15, WLC18 and WLQ20) did not load as expected and were removed. A principal factor analysis with a direct oblimin rotation was carried out again. Three factors were extracted, namely, calling, career and job orientations to work.

A principal component analysis was carried out on the 10 items of the WAMI. The eigenvalues of the first three factors and percentages of variance extracted were as follows: Factor 1 = 5.22 (52.15%); Factor 2 = 1.13 (11.29%) and Factor 3 = 0.84 (8.36%). Given that a three-factor structure was expected for the WAMI, and because Factor 3 explained 8.36% of the total variance, it was decided to retain the three-factor structure. Next, a principal factor analysis with a direct oblimin rotation was carried out. The three factors that were extracted were labelled as meaning making, greater good motivations and positive meaning. The items that loaded on the three factors that constitute meaningful work are in line with the factors identified by Steger et al. (2012). However, one item, namely, WAMI4 (‘I understand how my work contributes to my life’s meaning’) loaded on Factor 1 (meaning making) rather than Factor 3 (positive meaning), where it is supposed to load. Given that the item concerns work as a form of meaning making, it was decided to retain it on Factor 1.

A principal component analysis was carried out on the nine items of the WES. Two factors had eigenvalues larger than one. Factor 1 had an eigenvalue of 4.90 and explained 54.47% of the total variance. Factor 2 had an eigenvalue of 1.05 and explained 11.63% of the variance. Given that one factor of work engagement has been reported consistently in South African studies (see Rothmann, 2017), it was decided to retain only one factor. The component loadings ranged from 0.47 to 0.89, whereas communalities ranged from 0.70 to 0.86. The factor was labelled work engagement.