Keywords

1 Introduction

In the previous two chapters, the empirical results of the work identity research project were presented (Chap. 7) and discussed relative to existing literature (Chap. 8), respectively. This final chapter focuses on three sections, namely, the conclusions, the implications and recommendations for practice and the suggestions for future research.

  • The first section provides a broad overview of the most important conclusions of the research project emanating from the literature/theoretical review, the qualitative and the quantitative phases of the research project.

  • The second section then highlights the implications and practice-based recommendations on these aforementioned findings and conclusions.

  • The third and final section presents some suggestions for future research. These specific suggestions are grouped under three different headings.

2 Conclusions on the Findings

The main conclusions drawn from the findings of the work identity project (as reported in more detail in Chaps. 7 and 8) are the following:

Work identity is the answer to the question ‘Who am I at work?’ Work-based identity (WI) plays a central role in employee behaviour at work but more specifically in describing or explaining engaging behaviours at work. An overview of the main findings of the project will be provided briefly according to the different parts and chapters of this book: Chap. 1 explains the context and setting of the work identity project and explicates the different research aims of the project. Part I provides the conceptualisation and theoretical foundations of WI used in this book and consists of the following four chapters: Chap. 2 explicates the theoretical foundations for the conceptualisation and operationalisation of WI. Chapter 3 is based on qualitative empirical research and provides a model for explaining identity work and also the tactics and strategies applied to maintain and enhance WI. Chapter 4 focuses on an overview of the existing research literature with respect to the job demands and job resources as possible antecedents of WI. Chapter 5 shifts the focus of the literature review to the subjective and the objective consequences of WI. Part II consists of one chapter: Chap. 6 provides the exploratory empirical research findings of the two SA quantitative studies of the project. Part III also consists of one chapter that focuses on the testing for causality. Part IV, the last part, consists of two chapters: Chap. 8 provides a discussion of these findings indicating the specific contributions of this research project. Conclusions relating to the different chapters will be provided in greater detail in the following paragraphs below.

In retrospect, the decision to ground WI in social identity theory (SIT) (cf. Tajfel and Turner 1985) and role identity theory (RIT) (cf. Stryker and Burke 2000) that respectively emanate from sociology and psychology was a well-considered decision. SIT and RIT with their respective subcategories provide a sound theoretical foundation for integrating work-related commitment research. Subsequently, SIT and RIT provide the potential for integrating a wide array of seemingly disparate and unrelated commitment-related constructs to explain the relationship between individuals and their work or the work facets they identify with. These two broad theoretical streams (as described in Chap. 2) provide the so much needed theoretical foundation for work-related commitment research which is seemingly lacking in some of the current research streams on attachment, involvement, engagement and commitment to work.

WI remains a dynamic construct which is difficult to conceptualise and to capture operationally, especially the dynamic component of identity work. WI and identity work are two distinct but inseparable constructs. Work identity can be viewed as a dynamic condition or state (as an outcome), and identity work is the process leading up to this condition or state (as described in Chap. 3) (Adams and Crafford 2012). But this is a constant and a dynamic process, where the process influences the condition and that feeds back into the process again – in other words a double helix process that continues to develop in a reciprocal and a cyclical manner. This conclusion was again affirmed in the qualitative phase of the work identity project and is illustrated in the Saayman and Crafford (2011) model proposed in Chap. 3. Furthermore, the qualitative phase of this research project also identified nine different tactics and strategies that employees are using (in their identity work) to promote and maintain their work identities (refer to Chap. 3). These tactics and strategies may appear superficial at a first glance, but they suggest a broad taxonomy of tactics and strategies that are located in different segments/zones of the model suggested by Saayman and Crafford (2011) which can be further explored on deeper levels as was suggested in Chap. 3.

The distinction between antecedents and consequences of WI enabled the researchers to separate the WI construct from its causes and also from its consequences, thereby assisting in conceptualising and operationalising a clearer conceived construct that is less affected by construct contamination. The proposed onion model led Lloyd et al. (2011) to conclude that WI is a multilayered and a multidimensional construct, and their definition was used as a conceptual basis for operationalising the WI construct. Schaufeli and Bakker (2001) argued that employees’ psychological identification with work happens within four broad focal areas, namely, with work in general (work centrality/work involvement); secondly, with what the employee does (career/occupational/professional identity); thirdly, with where the employee works (organisational membership) and lastly, with the specific work the employee does (job involvement). The decision to also superimpose these different work-related identification foci on Kirpal’s (2004b) three-dimensional model (structural, social and individual-psychological) in the conceptualisation and construction of the WI construct and the WI prototype (refer to Chap. 2) yielded usable and interpretable results in both the Bothma (2011) and the De Braine (2012) studies (refer to Chap. 6). WI can therefore be viewed as an umbrella term that covers a range of different work-related identification foci – i.e. the social foci that individuals are identifying within their workplace. These may include identification foci such as the organisation, occupation/profession, career, work team, job or work role of the individual. An abundance of scales currently exists that can measure attachment, involvement, commitment or engagement in these different identification foci (refer to Roodt 2004).

The initial six theoretical dimensions used for the construction of the WI scale yielded three different interpretable dimensions in the Bester (2012) study (also refer to Chap. 7), namely, work centrality, person–organisation fit and value congruence. These three dimensions functioned effectively in the SEM to predict work engagement and turnover intentions in Bester’s study. The SEM clearly established some causes and consequences of WI.

The theoretical foundations and the process followed in the conceptualisation and operationalisation of the construct are fully explained in Chaps. 2 and 6. Chapter 4 provided an overview of the literature in respect of the traditional JD-R model and some additional selected job demands and resources and their relation to WI. Literature that specifically links job demands and job resources to WI is scarce. Even sparser is literature on empirical studies that examined the moderating effect of biographical and demographical variables on the relationship between job demands and resources with WI.

Chapter 5 also provided an overview of the literature in respect of the subjective (personal alienation, helping behaviours, work engagement and burnout) and the objective consequences (task performance and turnover intention) of WI and also indicated how these relationships were moderated by biographical and demographical variables. Research in this regard reports on how some work identity facets relate to these consequences, but not on a combined or integrated WI construct. There is also a clear void in the literature in this respect.

With reference to the second part of Chap. 6, it was established in the De Braine (2012) study that job demands (JDs) and job resources (JRs) (as in the JDRS model of Jackson and Rothmann (2005)) and other selected JDs (push factors) and JRs (pull factors) are predictors of WI. The term push factors is used to describe variables outside the traditional JD-R model that may in a similar way as JDs also lessen or decrease WI. On the other hand, pull factors similar to JRs may result in increased or enhanced WI. More specifically, it was found that the following JRs related to WI in a linear and in a non-linear way: growth opportunities, perceived external prestige, organisational support, team climate and advancement. Only one JD, namely, overload, was related to WI in a linear as well as in a non-linear way. (Refer to De Braine (2012) for a more detailed discussion of these non-linear results.) By comparing these linear and non-linear relationships between the JDs, JRs and WI, it was found that these differences were ignorably small. Both these sets of findings clearly established that the antecedent conditions (with reference to the JD-R model) are co-determining WI. Bester’s (2012) study (refer to the first part of Chap. 7) also found that different JDs (push factors) and JRs (pull factors) (outside the traditional JD-R model) are related to specific WI dimensions. More specifically, it was found that the following JRs predict WI: need for organisational identification, organisational reputation, task resources and remuneration perceptions. The two JDs that predicted WI were breach of psychological contract and work–family conflict. Both these JRs and JDs fall outside the traditional JD-R model. It can therefore be concluded that the mentioned JDs and JRs are predictors of WI and significantly contribute to the formation of WI. What is also evident is that the traditional JD-R model (discussed in Chap. 6) is not comprehensive enough to contain all possible JDs (push factors) and JRs (pull factors). The model can therefore be enriched by way of including other push and pull factors. All possible interaction effects between push and pull factors or JRs and JDs were not fully explored in this research project. Practitioners and scholars of WI should be cautioned not to focus only on linear cause-and-effect relationships but also on the interactive and reciprocal relationships when predicting WI. This can be more effectively dealt within SEMs or in cross-lagged panel design studies.

With reference to the third part of Chap. 6, it was also determined in the Bothma (2011) study that WI was significantly related to subjective (or more proximate) consequences such as personal alienation, helping behaviours, work engagement and burnout as well as objective (or more distal) consequences such as task performance and turnover intentions. In the case of the latter, turnover intention scores were used as a proxy for actual turnover (something more about this right at the end of this section). Findings of the study also indicate that the subjective (proximate) consequences mediate the relationship between WI and the objective (distal) consequences. These findings indicate that WI can be used to successfully predict important outcomes in the workplace such as task performance and actual turnover. Bester’s study (refer to Chap. 7) not only found that WI could predict turnover intention but that it could also predict work engagement. Based on these findings, the WI construct can be clearly distinguished from the work engagement construct, despite the fact that they are significantly, positively correlated and despite the fact that they share some degree of common variance. Tests for multicollinearity have established that these two constructs were significantly different.

From a methodological perspective, the decision to approach this work identity projectFootnote 1 in different phases was a constructive and productive decision. First, it helped by conducting a comprehensive literature review as the first stage and thereby developing a better understanding of WI and its underpinning theoretical foundations.

Second, it helped by conducting the qualitative phase of the empirical research first and thereby assisted the project team members in developing an appreciation for the construct in a natural research setting. As explained in Chap. 1, all the researchers in the project were initially involved in the qualitative phase of the study and could experience the ‘feel’ of the WI construct as perceived and experienced by research participants in this phase. This enabled them (research team members) to better understand the concept and its operationalisation.

Third, against the background of the first two phases, these two earlier phases assisted in conceptualising and operationalising the WI construct that was used in the third empirical, quantitative phase of the project (as reported in Chaps. 6 and 7). Researchers in this phase of the project could therefore develop a better idea about the foundational constructs and could qualitatively experience these in their respective research settings.

The conceptual (bow tie) model proposed in Chap. 1 (refer to Fig. 1.2 and Table 1.1) assisted in conceptualising a range of different causes and outcomes of WI. It was postulated in this research project that the antecedents of WI were significantly related to WI. This finding therefore clearly indicated that WI (as a criterion) could be successfully predicted and is partially co-determined by these selected antecedents. The postulated consequences in this project were also significantly related to WI (as a predictor), thereby indicating that WI can be successfully used as a predictor of WI consequences. The proposed research model provides a clear research agenda which can be used to systematically address research questions in the future.

With the construct-descriptive and the criterion-predictive validity firmly established in both the Bothma (2011) and De Braine (2012) studies (refer to both the second and third part of Chap. 6), the WI construct was also used in a structural equation modelling study (Bester, 2012) where both antecedents and outcomes were included in a predictive model (refer to Chap. 7). These findings clearly illustrate that the WI scale can be effectively used to measure the WI construct.

A finding not related to WI per se, but more specifically to the turnover intention scale (TIS) used to predict the consequences in the Bothma study, suggests that the TIS can be used as a proxy for actual turnover, since scores on this scale were significantly related to actual turnover after a 4-month and a 4-year period after the survey was conducted. This was a unique way of establishing the criterion validity of the TIS used as a consequence of WI in this research project. Tracking or tracing WI scores and turnover intention scores over time does provide useful indicators of employees’ attachment levels at work.

3 Implications and Recommendations

The practical implications and recommendations of this research project are based on these above-mentioned findings and conclusions. The practical implications are the following:

This research project has implications on different organisational levels in practice. More specifically, implications can be mentioned for policymakers on a labour policy level, for strategic HR managers on a company strategy level, for HR practitioners and HR consultants on a tactical level and for people supervisors and managers on an operational level. From a macro policymaking point of view (specifically relevant to the SA context), it should be noted that SA is urgently lacking a national work identity framework. SA labour policy should therefore be aimed at developing an enabling labour context. Such a labour context is required which provides the national resources or pull factors, e.g. a dedicated budget available for skills development, a national culture of skills development, a range of educational programmes in the workplace and the many others, too many to mention. Subsequently, these national resources or pull factors could facilitate the development of a basic work identity for all SA employees on all levels. Furthermore, the SA labour policy should ideally also address the most important constraints or push factors in the SA workplace, e.g. ignorance, the dangers of thinking in terms of ideology, violence, racism, discrimination and also many others that are too much to mention. These constraints or push factors may negatively impact on the development of a basic SA work identity. Such a national or a uniform WI framework constituted by the different push and pull factors may foster a greater understanding between different racial and cultural groups belonging to a very diverse workforce (as explained in more detail in Chap. 1). The authors are of the opinion that these principles may apply equally to other developing or culturally diverse countries that are facing similar development challenges.

On a company strategic (or meso) level, the HR managers concerned with strategic HR issues need to take note of which factors in the traditional JDRS (Jackson and Rothmann 2005) model (job resources) contribute to activating WI and which factors (job demands) contribute to deactivating WI. They should also understand how these forces interact with one another and how this impacts the formation of WI. A strategic culture change to implement and promote engaging work practices and an engaging workforce may have a significant effect on employee and ultimately company performance (refer to Chap. 8 for a discussion on WI outcomes).

On a tactical and a micro operational level, HR and line managers should be sufficiently informed and skilled in understanding how to develop and create a work context with which employees can and would like to identify with. What job resources (or pull factors) can be provided to grow WI, and what job demands (or push factors) should be restricted to create a conducive and an engaging work context?

The practical considerations presented above suggest that leaders/managers of unions and business organisations in all the different levels should be sufficiently informed about the dynamics of creating and promoting job conditions that would establish and grow a national WI (a work identity that applies to SA nationwide) in SA organisations or in other similar developing country contexts. (The conditions that contribute to the low WI levels were identified and discussed in Chap. 1.) More specifically, could SA leaders/managers give attention to the following job resources and demands (in this particular order) that partially co-determine the construction of WI?

  • Growth opportunities – require a work context that offers opportunities for growth and development

  • Perceived external prestige (organisational reputation) – aspires to be associated with an organisation that has a clear identity and a well-developed brand

  • Organisational support – prefers to work in an organisation that provides comprehensive support

  • Team climate – wishes to work in a supportive team environment

  • Advancement – prefers to see opportunities for advancement

The job demand that hampers the development of WI is:

  • Overload – requires some degree of workload that will pose some challenge to job incumbents

More specifically, in the SA context, the WI framework can be used to develop a unique SA work identity. Given the challenges and tensions that the SA society is facing today (refer to Chap. 1), this framework may assist in creating a more conducive work context for all SA employees and also a so much needed shared frame of reference when it comes to work-related issues and challenges. However, not all employees are similar, and practitioners should therefore be cautioned to allow for individual differences and preferences – since efforts to enforce a uniform approach in the extreme might lead to withdrawal and deactivation of work identities.

The WI construct provides potential parsimony in commitment-related research that is currently characterised by concept redundancy as well as construct contamination (refer to par. 9.2 on this point). The role that the work-related self-concept plays in integrating these different and often redundant work-related identification foci is currently under-explored and provides the potential for further empirical studies.

The research findings clearly show that WI is a construct that holds significant methodological potential. It provides a more parsimonious approach towards explaining the identification process at work, incorporating a host of other identification constructs and foci. It could therefore assist in bringing some degree of parsimony and order to the field of work-related commitment research and thereby address to some extent the concept redundancy and contamination issues. Some of these issues are explained in more detail below.

Based on the theorisation presented in Chap. 2, the authors of this book argue that WI is anchored in a sound and solid theoretical framework, as opposed to many other engagement and commitment constructs that evolved from practice without any seemingly (or at least without a sound) underpinning or guiding theoretical foundation.

The WI research project has more specifically shown that the JD-R model can also be used to effectively predict WI. The ratio and all possible variations of interactions between JDs and JRs also play, in the case of WI, a significant role in explaining the process of identification with work. The research findings also indicate that the traditional JD-R model alone is not sufficient in explaining WI, probably because it was not designed to do so in the first place. The model can however be enriched by way of including other nontraditional JDs (push factors) and JRs (pull factors). Using a force-field analysis methodology can assist in developing a more inclusive model.

The De Braine (2012) study (refer to the second part of Chap. 6) reported no interaction effects between JRs and JDs when predicting WI. This is most likely a company-, a sample- or a criterion-specific finding where there are no serious shortages of resources and where demands were probably perceived as challenges. Consequently, it should then also be established what interactions are present between JRs and JDs when predicting WI. Previous research on work engagement (note: not WI) clearly established such moderation effects. However, given the relatively limited number of JRs and JDs explored in the current project, it may also be the case with WI or even in other work contexts outside SA as was confirmed in the Bester (2012) study.

WI also has particular consequences that are either subjective or objective. The more proximate (closer to the individual) subjective consequences (personal alienation, helping behaviours, burnout and work engagement) were all significantly related to WI, and these also played a mediating role in the relationship with the more distal (further removed from the individual) consequences, namely, turnover intentions, work engagement and task performance. This means that distal consequences can be changed by giving attention to how the proximate conditions are perceived by employees.

Work engagement literature per se indicates a relationship between work engagement and organisational commitment. These findings suggest that some gliding scale of deepening identification exists ranging from identification, involvement, engagement to commitment. Such findings will enable practitioners to better understand the causes and the consequences of WI and beg for further investigation. Subsequently, the conditions leading to high WI can be more effectively modelled and simulated.

If practitioners know what the antecedents and consequences of WI are, they can either enhance conditions that activate WI or reduce those that deactivate WI. But since WI formation is a highly individualised process, practitioners need to keep individual preferences and differences in mind. Negative consequences of WI can thereby be indirectly manipulated or reduced, and positive outcomes indirectly created or enhanced with sufficient knowledge of individual preferences. The proposed bow tie model (refer to Chap. 1) can provide a framework for understanding the causes and outcomes of WI.

WI is a relatively ‘new’ construct in the research literature. If JDs and JRs (in the broadest sense) co-determine WI, it should be explored in practice which JRs relate more strongly to the activation of WI than others. The opposite is also true; it should be further investigated and established which JDs lead to the deactivation of WI as opposed to others. Thirdly, it should also be determined how these JDs and JRs interact amongst themselves and in relation to WI. This information would enable managers and supervisors to create optimal working conditions for modern-day workers. The JDs and JRs in the current study were mainly linked to selected contextual resources and demands based on the JDRS model which is clearly not all inclusive. Evidence suggests that the model can be further enriched.

WI as a state or as a condition is conceptualised and operationalised in such a way (refer to Chaps. 2 and 4 for a detailed discussion) that it can be distinguished from its antecedent conditions as well as its consequences. Such an approach will ensure that the essence of WI is captured more accurately, compared to other measures that erroneously also include causes or consequences of engaging behaviours at work. Following this approach in the current study addressed the issue of possible construct contamination between WI and its causes and consequences. The value of conceptually separating the WI condition from its causes or consequences lies in the consistency in which WI can be measured despite shifts or changes in the contextual factors. Stated differently, this means that if the contextual factors are changing, it will only result in a change in the level of WI scores and not in the nature of the WI construct itself.

4 Suggestions for Future Research

Suggestions for future research are grouped under the following three subheadings:

4.1 Further Exploring and Expanding the Research Model

4.1.1 Exploring Other Variables

The bow tie model of antecedents and consequences of WI proposed in Chap. 1 can be used to systematically investigate the causes and consequences of WI. The list of all possible JDs and JRs or other push and pull factors (not only those included in the traditional JD-R or JDRS models) is therefore not fully explored. The inclusion of other demands and resources would enrich the predictive model and may explain additional variance in the prediction of WI. On the other hand, it may be possible that a totally different set of push and pull factors may operate to activate and optimise the process of identity work (IW). It is suggested that the complete list of potential JDs and JRs as well as other push and pull factors are explored in future studies, so that it can be determined which JDs and JRs as well as other push and pull factors best predict WI or IW. The continuous double helix process between WI and IW should also be considered in this model. It should also be realised that this list will never be ‘complete’ in the true sense of the word because it will be continuously ‘updated’ due to disruptive transitions in the twenty-first-century workplace. The Bester (2012) study has clearly indicated that other variables outside the traditional JD-R model also play a significant role in WI formation. The JD-R model should therefore be enriched or entirely reconceptualised in the explanation of WI per se.

4.1.2 Including Personal Resources

We furthermore suggest that personal resources should be included in further predictive research on WI, specifically those included in the psychological capital (PsyCap) model as suggested by Luthans and research associates (cf. Luthans and Youssef 2004; Luthans et al. 2007). These personal resources may explain more variance in WI than contextual factors (e.g. job demands and job resources or push and pull factors) alone. It is also postulated that there may be interaction (mediation) effects between these contextual push and pull factors and the PsyCap elements (hope, efficacy, resilience and optimism), possibly resulting in more variance explained in the prediction of WI. This line of reasoning is based on the fact that PsyCap elements are more closely related to the individual’s self-concept and will as such influence WI formation more closely.

4.1.3 Activation or Deactivation of WI

It would also be interesting to explore the ratio (relative strength) between specific JD and JR scores to determine under which conditions WI on a gliding scale is more activated or when WI is less or even deactivated. It would then be possible to determine which ones are the most important JRs (pull factors) that activate WI and which ones are the most important JDs (push factors) that deactivate WI. It may also be that the salience of specific individual needs is related to the JRs (or pull factors) or inversely to JDs (or push factors) and that these may activate or deactivate particular work identity facets. It is also unknown at this point in time how these push and pull factors interact amongst themselves and with changes in different working conditions. Stated differently, the question would be if the push and pull factor relationships with WI would remain constant in different working conditions and contexts. These relationships need to be explored and investigated further.

4.2 Conceptual and Operational (Theoretical) Matters

4.2.1 Exploring the Work Identity Construct

More qualitative research should be conducted on the WI phenomenon – more specifically on the link between WI and the process of identity work. At the very start of this research project, WI was viewed as a ‘black box’ concept. Though we could unveil and conceptualise WI to some extent in this research project, WI still remains to some degree at this point in time to be still considered a ‘black box’ concept – something we don’t fully grasp or that we can’t fully explain yet. Through qualitative research an appreciation can be developed in how the identity formation process unfolds and what particular stages are contained in the process of resolving the tensions between the personal and social-based work identity foci. These tactics and strategies that employees apply at work have not been fully explored yet, and the proposed taxonomy in Chap. 3 suggests a more systematic approach towards investigating possible strategies and tactics in different domains of the proposed model. The integrative role of the self-concept in resolving tensions between conflicting work-related identification foci is currently under-explored. We recommend that the integrative role of the self-concept during these tension-resolving processes should be further explored.

Closely related to the above is the issue relating to different modes of individual–organisation bonding process. In some instances individuals have a personal relationship with their organisations or their representatives. In other cases, the relationship is mediated by a third party such as a trade union or a trade union representative. These two processes follow distinctly different routes of socialising newcomers, and the question is whether this results in distinctly different work identities.

It is further recommended to explore this field in more depth in an interdisciplinary research setting of psychology, cultural/social sciences, business administration and HRM.

4.2.2 Including Affective or Emotive Components

The fact that people are or become passionate about their work raises the question that WI may also contain affective/emotive and intentional elements (not only a cognitive element). Consequently, the question can be raised: what exemplary behaviours are related to these other two elements that would better describe the WI construct? Further qualitative and quantitative enquiry may shed more light on the WI construct that may contain cognitive, affective and conative elements but that would still be distinguishable from other constructs such as job satisfaction (that is measuring affect in the workplace) or other workplace intentions such as turnover intentions (that measures intentions to leave or stay).

4.2.3 The Identification Continuum

As suggested in the earlier chapters (refer to Chaps. 2 and 6), levels of WI may exist on a continuum with under-identification and over-identification on the extreme ends. If WI is viewed as a bipolar continuum, is there a shift/change in the nature or composition of the identification construct if one moves from one position to another position on this suggested continuum? A related issue may be how this WI continuum and its extreme poles are related to certain outcome variables on a gliding scale such as well-being and psychological health on one end or related to alienation, burnout, type A behaviour and workaholism on the other end?

4.2.4 Work-Identification Facets

It would be interesting to know what the salient work facets (identification foci) are that individuals prefer to identify with. How do these work facets relate to individual need salience – in other words, is there a specific relationship between particular individual need salience and the choice of work facets individuals like to identify with? Is the relationship between particular individual needs and WI possibly mediated by specific job resources and job demands? Then, is this choice of particular work identity facets culturally/socially imposed? This knowledge may explain differences in social or cultural preferences in WI – a suggestion made by Bester (2012) in his study.

Closely related to the above is an issue associated with culturally diverse work settings. If different cultures work in the same organisation, do these cultures bring different social identities to the workplace? How would these different social identities then affect the work identity formation process? Do different cultures give preference (or value) to different identification foci in the workplace? All these questions are still unresolved and beg for further investigation.

4.3 Methodological Suggestions

4.3.1 Exploring Different Research Settings

From a methodological perspective, WI was only investigated in one specific setting – namely, the ICT sector in SA – in this research project which has limit generalisation value. It would be productive to also expand this SA project to other organisational, industry or country settings to test if the research model holds across different contexts and if these findings could be generalised across different organisational, industry or national boundaries. This is an aspect often neglected in published research in general (Roe 2012; Roodt 2012).

Closely related to the above-mentioned aspect is the term ‘difficult’ working conditions. How does working in difficult conditions, such as underground mining (extreme temperatures and high humidity) or in call centre environments (sometimes referred to as ‘sweat shops’), affect the identity formation process at work? Does identity formation evolve in a different manner compared to ‘normal’ working conditions? These questions beg further research.

4.3.2 Cross-Lagged Panel Design

As suggested earlier, it appears that work identification (work identity), job involvement, work engagement and organisational commitment are independent but related constructs. It would therefore make sense to test the causal and interactive sequence of these (and other related) constructs in a longitudinal, cross-lagged panel design study where the double helix dynamics can also be considered. It would also indicate what common variances are shared between these constructs. This information would provide answers to the questions in what sequence or interactive relationships these constructs are activated and which ones are triggering or causing the others. Such a study may cast a totally different light on how the work identification process starts and develops across different possible phases or stages. It is postulated that different but deepening levels of identification exist across such stages. These identification stages may possibly be linked to different organisational tenure or career development stages.

4.3.3 Non-linear Relationships

Preliminary research findings indicate that there may be non-linear relationships between JRs and WI as well as between JDs and WI (refer to Chaps. 7 and 8). These findings suggest that there is an optimal point of a JR or a JD after which WI will increase or decrease (depending on whether it is a U or an inverted U relationship). The existence of such non-linear relationships should be further explored, since it is unknown whether it is a resource-specific or a demand-specific phenomenon or whether it applies to JRs or JDs (or push and pull factors) in general.

4.3.4 Including Objective Measures

On a more practical note, since the work identity project mostly made use of self-report measures, it is suggested that more objective outcome measures (i.e. not only self-report measures but also others such as financial performance of work units, days absent from work, productivity measures, etc.) are also included as dependent variables in future studies.

4.3.5 Translation Effects

English was used as the medium in the surveys that were conducted during this research project. This is to most of the research participants in the SA context their second or third language. English language proficiency is therefore a crucial factor to be considered when conducting survey research in a multicultural research setting. At the same time, we realise that using one single language – in this case English – may filter out some relevant monolingual, local-language-speaking participants. However, in an electronic research setting, different languages can be used simultaneously. Translations or back-translations of surveys immediately open up a different set of issues that are related to measurement equivalence – an issue that was already alluded to earlier.

4.3.6 Measurement Equivalence

In order to establish the measurement equivalence of the WI scale, it is perhaps self-evident to test if the scale is equivalent for different race, ethnic/language as well as gender groups. Such analyses on potential bias were not conducted on the WI scale during this research project. The range of nonsignificant findings on the moderating relationship of biographical and demographical variables and WI suggested that such analyses were not necessary.

5 Concluding Remarks

The work identity research project reported in this book tried to answer the question ‘Who am I at work?’ within the dynamic and changing SA work context. In order to address this question, the researchers approached the problem in two distinct empirical research phases. The first phase was aimed at establishing how work identity is formed, developed and maintained from a qualitative research methodology perspective. With this phase concluded, the researchers then engaged in a quantitative research mode in the second phase to establish how WI can be operationalised as a predictor or as a criterion. This was done more specifically with a study on the prediction of WI by using the JD-R model as well as a study on the consequence of WI that investigated subjective and objective outcomes. A third study in this second phase developed and tested a structural equation model where WI played a central role in the proposed model. Results of these studies established that WI is a construct worthwhile to pursue (even in multicultural work settings). Because the WI construct can be applied as a predictor and also be related to important consequences that are in turn linked to individual and indirectly to organisational performance, the research objectives of the work identity research project were hereby achieved.