The overarching objective of this study is to develop a first version of a CSR leadership questionnaire (CSR-LQ) that would serve as basis for developing a CSR leadership competency model in future. Specific objectives are firstly to derive CSR leadership behaviours or competencies from available literature on leadership and CSR and secondly to develop a first version of a CSR leadership measure that could form the basis of further empirical testing to be ultimately used to obtain 360° assessments of the CSR leadership competencies.

In the literature, a wide variety of definitions of CSR are reported. The multilevel and multidisciplinary nature of CSR including the predictors, outcomes, mediators, and moderators are extensively discussed by Aguinis and Glavas (2012). For succinctness, the authors will thus only highlight the definition of CSR that was used as the basis for scale development in this study.

Carroll (1991) created a four-part definition that defines CSR as follows:

The social responsibility of business encompasses the economic, legal, ethical, and discretionary (later referred to as voluntary) expectations that society has on organisations at a given point in time. (p. 283)

This definition is useful in that it specifies the organisation’s economic responsibility as a factor to be considered in CSR and viewing an organisation’s corporate responsibility as inseparable from its economic function (Carroll & Shabana, 2010; Inyang, Awa & Enuoh, 2011). The socially responsible business can thus be viewed as one that attempts to minimise its negative effects on society, whilst at the same time utilising its resources in such a way to maximise its positive effects.

The economic responsibilities of a business include producing goods and services that society wants, at a price that maintains the business, whilst at the same time satisfying its obligations to its shareholders (Carroll & Shabana, 2010; Matten, Crane & Chapple, 2003). Legal responsibilities refer to the positive and negative obligations businesses have by law and through regulations in the society in which it operates. Thus, it refers to obeying the laws and regulations of society (Carroll & Shabana, 2010; Ferrell et al., 2011; Van den Ende, 2004). Ethical responsibilities comprise meeting other societal expectations not written in the law, whilst voluntary responsibilities (philanthropic) are additional behaviours and actions that society finds desirable and which complement the business’s values (Ferrell et al., 2011).

Carroll (1991) developed the notion of CSR more explicitly when he contended that the economic and legal responsibilities of an organisation can be considered as those that society requires from them, the ethical responsibilities are those that society expects from them and the voluntary responsibilities are those that society desires from them. A distinction was thus made between the traditional responsibilities of organisations, which are embodied in their economic and legal responsibilities towards society, and the new responsibilities of organisations, which are embodied in their ethical and voluntary responsibilities and reflect the new, broader, social contract between organisations and society (Carroll & Shabana, 2010). For the purpose of this study and the development of a CSR leadership measurement instrument, Carroll’s four-part definition of CSR is used.

After a comprehensive leadership literature review covering the trait theories, behavioural theories, contingency theories and the neo-charismatic theories, it was evident that the theories of the neo-charismatic leadership paradigm are especially useful when identifying the competencies that are necessary to be a successful CSR leader, as the focus is on the behaviours that constitute outstanding leadership (Robbins, Odendaal & Roodt, 2003).

Strand (2011) carried out a review of existing research relating leadership to CSR with the primary goal to try and better understand the intersection between the two fields and to identify potentially fruitful avenues of future research. His review indicates that a number of researchers have explored the relationship between leadership and CSR in terms of the neo-charismatic leadership theories, including transformational leadership (e.g. Egri & Herman, 2000; Lacerda, 2010; Shahin & Zairi, 2007; Waldman, Siegel & Javidan, 2006, as cited in Strand, 2011), transactional leadership (e.g. Szekely & Knirsch, 2005, as cited in Strand, 2011) and visionary leadership (e.g. Waldman, De Luque, Washburn & House, 2006; Ketola, 2006, as cited in Strand, 2011). Strand noted that Waldman, Siegel and Javidan’s (2006) exploration of transformational leadership factors and how they relate to CSR activities ‘represents one of the more thorough, descriptive studies’ (p. 91) regarding the intersection of the two fields.

In 2006, Waldman et al. investigated the potential link between CEO transformational leadership factors in 112 large US and Canadian companies and the CSR actions of their respective firms. Citing an absence of independence between the two emotional factors idealised influence and inspirational motivation (Waldman et al., 2006, p. 1706), they examined both as a single emotional factor termed ‘charismatic leadership’. The intellectual factor intellectual stimulation was also examined, but the factor individualised consideration was excluded due to conceptual complications in connecting individual-level focus with the higher-level organisational occurrence of CSR.

Waldman et al. (2006) report a positive relationship between CEO intellectual stimulation and the firm’s tendency to engage in strategically-oriented CSR, but not socially-oriented CSR. According to Waldman et al., the results suggest that intellectually stimulating leaders not only attempt to practise corporately responsible actions, but also aim their efforts at areas that are most pertinent to the firm’s strategic objectives. However, in contrast to what was hypothesised, Waldman et al. found no statistically significant relationship between CEO charismatic leadership and CSR, proposing that the laudable features of charismatic leadership that are emphasised in the literature may not actually lead to corporately responsible actions.

Building on the work of Waldman et al. (2006), Lacerda (2010) investigated the relationship between CEO transformational leadership and CSR in 50 large Portuguese firms representing different sectors. Lacerda examined all four transformational leadership dimensions as outlined by Bass and Avolio (2008), as well as an additional dimension ethical integrity as described by Turner, Barling, Epitropaki, Butcher and Milner (2002). The study found that transformational leadership correlated positively with firm strategic orientation to CSR, with specifically the two dimensions inspirational motivation and individualised consideration strongly linked to firm strategic orientation to CSR. Idealised influence and intellectual stimulation were not found to be significant predictors of firm strategic orientation to CSR. Despite the theoretical background linking the two (Lacerda, 2010, p. 30), it was found that ethical integrity is not significantly related to firm strategic orientation to CSR, implicating that a leader’s ethical integrity is not significantly linked to the adoption of CSR practices (Lacerda, 2010).

In 2009, Kakabadse, Kakabadse and Lee-Davies conducted research that focused on both the leadership skills and capabilities needed to effectively implement CSR in organisations. They presented a model that included how to implement and drive CSR through organisations, outlining three stages of CSR implementation and identifying the different leadership skills that would be needed in each stage. The three stages of CSR implementation identified by Kakabadse, Kakabadse and Lee-Davies (2009) are the decision-making stage, the adoption stage and the commitment stage. In the first stage, the decision-making stage, the CSR leader comes to terms with the need for personal growth and organisational movement towards CSR. This is followed by the adoption stage, in which real movement begins and starts to impact the CSR leader’s organisation. Finally, the commitment stage is reached when it becomes clear that the organisation is committed to CSR in the long run and that CSR goals will be constantly pushed in spite of obstacles such as opposition by shareholders, complications in measuring CSR and conflicting priorities. Specific CSR leadership skills and capabilities needed in each of the three stages were also identified, each fulfilling its own role in the effort to achieve CSR and together signifying a coherent whole that offers the channel for organisations to accomplish CSR. In total 10 leadership capabilities were identified, namely CSR awareness and reflexivity in stage 1, discerning CSR goals, using business case language, persuasion, handling paradoxes and conflicting priorities, consistency of application, CSR measurement and follow-through in stage 2 and the will to act in stage 3. The three-stage model thus represents a journey to effectively implement CSR in organisations and is based on two important components, namely:

According to Kakabadse et al. (2009), all 10 leadership capabilities are interconnected and they often need to be used in conjunction with one another to face the unavoidable challenges confronting the implementation of CSR in organisations. Truly effective CSR leadership, which succeeds in fully implementing CSR in organisations, requires the responsible individual to use all the capabilities together, allowing the organisational leaders to act in a consistent way that makes sense to external stakeholders and drive the organisation forward. Kakabadse et al. came to the conclusion that CSR leadership can be developed, but that learning to master the skills needed to perform in each stage to successfully implement CSR in organisations can be challenging. They also acknowledge that observers may perceive their CSR leadership model as being borne of essentially charismatic leadership qualities that can be developed by individuals.

The available literature on the intersection between CSR and leadership thus shows that a number of researchers have explored the relationship between leadership and CSR in terms of the neo-charismatic leadership theories. The most thorough study is Waldman et al.’s (2006) exploration of transformational leadership factors and their relationship to CSR activities. This study concluded that transformational leadership is positively related to a firm strategic orientation to CSR. The studies by Waldman et al. and Lacerda (2010) therefore support the earlier notion that the theories of the neo-charismatic paradigm – and especially the transformational leadership theories – are useful when explaining the link between leadership and CSR and can be used as basis for identifying the CSR leadership competencies necessary to be an effective and successful CSR leader.

In 2002 Spangenberg and Theron developed a uniquely South African leadership questionnaire, the Leadership Behaviour Inventory (LBI) to assess, in terms of the South African organisational context, a wide range of capabilities required by leaders and managers to implement change whilst maintaining unit performance. This questionnaire incorporates numerous leadership theories (referred to earlier). Spangenberg and Theron (2002) argue that the leadership behaviours (competencies) identified by their model best describe the leadership behaviours that are necessary to be a successful and effective leader.

Spangenberg and Theron (2002) considered leadership effectiveness of critical importance whilst developing the LBI and felt that ‘leading change’ and ‘ensuring effective organisational unit performance’ were the most important aims of leadership in the South African organisational context (Spangenberg & Theron, 2002, p. 12). It was also postulated that the instrument should assess leadership and the change process in stages. Furthermore, the LBI should have the ability to measure the full range of behaviours required to successfully manage the change process, whilst reflecting the diversity of South Africa’s managerial population (Spangenberg & Theron, 2010).

The LBI is based on a comprehensive interpretation of the leadership construct, containing elements of leadership, management and supervision (Spangenberg & Theron, 2010). Charismatic or transformational leadership (general leadership) forms the core of the LBI leadership construct, whilst elements of management and supervisory leadership also make a significant contribution to the construct (Spangenberg & Theron, 2010).

Based on the compelling need for change and transformation in South Africa, Spangenberg and Theron (2002) thought it was important for the LBI to be based on theory from the neo-charismatic leadership paradigm. Eleven generic behaviours outlined by House (1995), which the author believed characterised outstanding leadership, formed the theoretical background of the LBI dimensions. A review of the literature also indicated that Conger and Kanungo’s (1987) model of charismatic leadership (as cited in Spangenberg & Theron, 2002), also known as the C-K model, conceptually best satisfied the demands of the South African context and served as the launching pad for the LBI development. The C-K model comprises three stages, namely the environmental assessment stage, the vision formulation stage and the implementation of vision stage (refer to Van Zyl, 2013, for a comprehensive discussion of the C-K model).

In addition to the three stages of the C-K model, Spangenberg and Theron (2010) decided to include an additional leadership role in their model due to the ‘absence of sufficient infrastructure in stage 3’ (p. 49). They argued that the group of behaviours categorised under empowerment was not adequate to guarantee that the desired changes would in fact be implemented. Nadler and Tushman (1990) echo the importance of infrastructure in implementing vision in complex organisations and suggest that instrumental leadership be utilised to supplement charismatic leadership. According to Nadler and Tushman, the aims of instrumental leadership include building competent teams, clarifying necessary behaviours, integrating measurement and administering rewards and punishments so that individuals view behaviours that are in line with the change as essential for them to fulfil their own goals. Furthermore, Nadler and Tushman point out that, in order to implement instrumental leadership, three elements of behaviour are required – namely structuring, controlling and rewarding – that are complementary to charismatic leadership (which includes envisioning, energising and enabling). However, Spangenberg and Theron (2002) point out that the success of any leader ‘will ultimately be evaluated in terms of unit performance that … cannot be achieved without attention to the visionary and motivational aspects of leadership, combined with instrumental leadership’ (p. 12).

The first generation LBI model comprised four phases measured by 24 dimensions (Spangenberg & Theron, 2002). After the instrument was extensively used and its psychometric properties evaluated, the original LBI was accordingly revised to a more streamlined, second-generation inventory (LBI-2) by reducing the four stages to three and by decreasing the number of dimensions from 24 to 20. The focus now turns to the research design followed in the current study.

The study made use of a mixed-method ex post facto research approach, which included both qualitative and quantitative phases. Netemeyer, Bearden and Sharma’s (2003) scale development framework was used as a point of departure for the development of the CSR-LQ and a three-phase research methodology was developed for this study as depicted in Figure 1.

During Phase 1 of the development of the CSR-LQ, CSR leadership as a construct was defined. This was done through conducting a literature review and making use of expert judging with the help of the basic principles of the critical incident technique (CIT).

After doing an extensive literature review, the LBI-2 was identified to form the basis for the development of the CSR-LQ. The dimensions of the LBI-2 were adapted for CSR leadership and served as starting point for the development of the CSR-LQ (see Table 1).

The adapted LBI-2 dimensions were subjected to expert judging by five CSR leaders as an aid for establishing face and content validity. Respondents were asked to indicate the relevance, specificity, representativeness and clarity of dimensions and no changes to the 20 dimensions were proposed (refer to results and discussion). The basic principles of the CIT developed by Flanagan (1954) were subsequently used to gain insight into CSR leadership and served as the basis for the generation of new items for the CSR-LQ (in Phase 2). This technique was especially useful in this study, as ‘critical incident’ refers to any behaviour that can be observed (Flanagan, 1954) and could therefore be used as basis for the development of new items for the CSR-LQ.

Over the years the CIT has advanced in two major ways. According to Butterfield, Borgen, Amundson and Maglio (2005), the CIT has firstly moved from being mainly behaviourally based to also include the study of psychological states or experiences. Secondly, there was a shift in emphasis by experts from direct observation to an incorporation of reflective self-report.

Although these two changes in the CIT have been shown to be fruitful in various research studies, it also brought along some challenges in the form of increased subjectivity during the gathering of data. The problem can be addressed through the implementation of credibility checks during the CIT process. Borgen, Amundson and Butterfield (2008) did a comprehensive review of the CIT literature and outlined five key steps and credibility checks (to minimise the effect of subjectivity of data gathering) in the CIT process.

Step 1: Ascertain the general aims of the activity to be studied. This step is closely linked to the research question, referring to the activity or construct that the researcher wishes to observe or have research participants self-report on.

Step 2: Make plans and set specifications. Four specifications need to be established and communicated before collecting the data, namely (1) defining the types of situations that need to be observed or reported on, (2) determining the situation’s relevance to the general aim of the study, (3) gathering a complete understanding of the extent of the situation’s effect on the general aim and (4) determining who will make the observations or take part in the self-reporting.

Step 3: Collect the data. The collection of data can be done by supervisors or experts in the particular field through the direct observation of people performing the activity or through subject experts recalling past incidents or experiences. The latter can be done through personal or telephone interviews or by making use of a questionnaire. For the purpose of this study, interviews were conducted with five CSR leaders to collect the initial data. For optimal data collection during such an interview, it is very important that a good relationship is established at the beginning of the interview. This will build trust between interviewer and interviewee, generally making the interviewee more comfortable to share information with the researcher (Miczo, 2003). After a relationship was established, the research participants were asked to recall the critical incidents that helped or hindered them in their role, the meaning of these incidents to them and the outcome of using the critical incidents by giving a practical example (if possible). This was done with the help of an interview schedule based on the CIT interview guidelines outlined by Latham and Wexley (1994) (refer to Addendum A in Van Zyl, 2013).

Step 4: Analyse the data. The process of analysing the data entails (1) determining the data’s frame of reference that surfaces as a result of its intended use, (2) formulating the different categories by means of an inductive process and (3) determining the degree of specificity or generality that is going to be used when reporting the data. Thus, a framework with different categories and operational definitions is built according to which the data can be organised for analysis.

Step 5: Interpret the data and report the results. A sequence of nine credibility checks can be used to establish the trustworthiness of the results of a CIT analysis:

The utilisation of these credibility checks ensures that the results obtained are accurate and a reflection of the job requirements or human behaviours critical for success in the occupational field being studied and positions them in the broader research and professional community (Borgen et al., 2008).

During the development of the first draft of the CSR-LQ, all the five steps were followed; however, only two of the checks listed above (numbers 1 and 5) were performed given the constraints of the study. This could be regarded as a limitation in the current study and is reported as such.

Purposive sampling was utilised to identify the five research participants (refer to Table 2) that participated in Phase 1 of the measurement instrument development process.

The second phase of the CSR-LQ development involved the generation of items for the questionnaire, as well as the expert judging of the pool of measurement items. There were various sources of items; the main source of items was the LBI-2. Minor adaptions to the LBI-2 items were made, based on the results obtained from the first phase of the scale development process, as well as from the literature on CSR and leadership.

Minghat, Yasin and Udin (2012) posit that the Delphi technique can essentially be described as the use of expert opinion, with the main objective of acquiring highly reliable responses towards problems and surveys given to a panel of experts. The authors acknowledge that they cannot claim that the Delphi technique was applied in accordance with all four guidelines as proposed by Su, Win, & Chung (2010), which include anonymity, iteration, controlled feedback and calculating each content validity index and the aggregation of group answers. However, some basic principles of the Delphi technique were used to judge the relevance of the newly adapted measurement items and to propose changes.

A panel comprising of 13 individuals (selected by means of purposive sampling and snowball sampling) who are experts in either the academic field of CSR, or who are active in the private sector as CSR leaders, participated in the judging of the CSR-LQ measurement instrument. The importance of the selection of experts is also emphasised by Su et al. (2010, p. 702) in that they state that ‘selecting “experts” as the participants of the Delphi group are very important because they need to understand and have experiences which are related to the specific research theme of performing such method’.

The pilot questionnaire was sent out to the selected respondents electronically. They were requested to indicate the importance of each item with regard to CSR leadership on a five-point Likert-type scale ranging from 1 (not at all important) to 5 (absolutely critical). Respondents were afforded the opportunity to give additional comments about specific items (e.g. appropriateness, domain representation, domain relevance, wording, grammar, clarity and comprehensiveness). The guideline of iteration (having respondents reconsider their responses at least twice to reach consensus) could not be met due to the sample constraints of the study.

After the successful completion of the first and second phases, the final version of the questionnaire was developed.

The literature review and expert review process culminated in the adaption of the LBI-2 for CSR leadership as presented in Table 1. It is argued that each of the leadership dimensions is relevant to CSR leadership and that the LBI-2 contains leadership competencies or behaviours that are also essential for successful and effective CSR leadership in organisations. In the following sections, arguments are presented as to why each CSR competency is a prerequisite to achieve the objectives of CSR. These arguments are essentially adaptations of the arguments posed by Spangenberg and Theron (2011) in the development of the LBI-2 and they serve to motivate the inclusion of each of the LBI-2 dimensions in the development of a CSR leadership measurement instrument.

In total, 119 items of the LBI-2 were revised and 33 new items were developed for the CSR-LQ pilot questionnaire. However, no LBI-2 items were adapted for the two CSR-LQ dimensions CSR leader self-discovery, reflection and self-awareness and displaying sound interpersonal skills, because the LBI-2 items represent generic leadership behaviours that are appropriate in different situations. The newly developed CSR-LQ items were classified according to the structure of the LBI-2 by adding the items to the corresponding dimensions. The rationale behind this decision was the fact that the LBI-2 comprises three actionable phases that are required to successfully implement a vision, which are also relevant to the implementation of a CSR vision.

A two-stream approach was used to improve or delete items, namely: (1) analysing comments on specific items to improve certain items or to delete items and (2) the use of descriptive statistics and numerical decision rules (refer to Table 3).

The comments received were extremely helpful in improving the content and face validity of the items. The comments were implemented by: (1) deleting items that respondents found troublesome or difficult to interpret, (2) merging similar items to create new items and (3) rewording items that were confusing to the respondents but did show promise. The descriptive statistics were interpreted to retain items that the expert judges felt are relevant to CSR leadership, whilst the means and standard deviations were used to identify those items that had high inter-judge agreement (refer to Table 3). In most cases, the comments received corresponded with the results of the descriptive statistics. Table 4 depicts how items for the CSR-LQ were modified and retained prior to and after expert judging. An X indicates inclusion of the item.

After the completion of the data analysis, the final version of the CSR-LQ was developed (refer to Table 4, last column). In total, 36 items were deleted from the pilot CSR-LQ questionnaire, whilst seven new items were added through the merging of items. The final version of the CSR-LQ hence consists of 123 items.

The main motivation for developing the CSR-LQ was to develop a measure of CSR leadership competence to enable the eventual development and empirical testing of a CSR leadership competency model. It is therefore recommended that the CSR-LQ measurement instrument firstly be revised once more using the critical incident and Delphi techniques according to the methodological rigour as proposed by scholars. Secondly, the CSR-LQ should be validated using structural equation modelling (SEM). The SEM process is based on covariances that are less stable when estimated from small samples and thus a large sample of CSR leaders with one or more followers would be needed. Once the SEM process has been completed and credible evidence on the reliability and construct validity of the instrument exists, the CSR-LQ can be used with confidence to operationalise the 20 latent behavioural CSR leadership dimensions as presented in the model. Once the CSR-LQ has been validated (meaning that the competencies of a successful CSR leader have been identified) future researchers can focus on identifying the outcomes of CSR leadership and continue with the development of a comprehensive CSR leadership competency model.

Since this study was the first step toward the development of a comprehensive, all-inclusive CSR leadership competency model, it is premature to draw conclusions regarding the instrument; conclusions can only be made after the CSR leadership measurement instrument that was developed undergoes a further phase of redevelopment and empirical testing. However, based on the data gathered from the 18 respondents (representing various industries’ representatives) during the development of the CSR-LQ, six tentative conclusions, with regard to CSR in general, can be drawn.

Firstly, the CEO, assisted by the group or company secretary, is primarily responsible for setting the organisation’s policy on CSR and providing oversight, whilst specific positions are then created to implement the organisation’s CSR strategy and to coordinate the various components of their CSR initiatives. Secondly, some South African companies are committed to CSR and realise its importance as a way to address South Africa’s societal and environmental issues. They also see the business benefits of engaging in CSR and view it as part of their competitive advantage. Thirdly, although the terms corporate social responsibility and corporate social investment have different meanings, the two terms are used interchangeably by many South African organisations. Fourthly, the release of the draft National Development Plan (NDP) entitled ‘Vision for 2030’ in November 2011 and the acceptance of the revised NDP 2030 entitled ‘Our future – make it work’ by the South African parliament in August 2012 impacted the CSR strategy and CSR initiatives of many South African companies. The NDP motivated companies to review and revise their CSR strategies and to align their CSR outcomes with the NDP in order to assist the country in achieving the plan’s stated objectives. Fifthly, many South African organisations align their CSR strategies with their core competencies and use their available resources to achieve their stated CSR outcomes and, lastly, in order to successfully implement CSR throughout the organisation, it is important to make it part of the organisation’s culture.