Locus of control and attitudes to working in virtual teams

INTERNATIONAL JOURNAL OF PROJECT MANAGEMENT International Journal of Project Management 24 (2006) 234–243 www.elsevier.com/locate/ijproman Locus of control and attitudes to working in virtual teams Liz Lee-Kelley * Strategic Change and Project Management, School of Management, University of Surrey, Guildford, Surrey, GU2 7XH, United Kingdom Received 5 August 2005; received in revised form 1 November 2005; accepted 20 January 2006 Abstract Global projects using new computing and communication technologies for at-distance communication and coordination are increasingly popular. As a result, team members are often located geographically and temporally apart, while working towards a common project goal. This paper takes a micro-level look at those actively involved in distributed team working. For the first time within the project team context, this paper uses the locus of control (LOC) construct to examine the influence of individual workers’ general control expectancies on their attitudes towards distributed working. Decisions and actions of management and team colleagues are perceived and interpreted by team individuals as either enabling or constraining their own performance. This in turn, can affect their behaviour towards others. With its basis in reinforcement or learning theory, LOC offers a robust theoretical framework to research the virtual project team and inform business practice. The paper first reports the results of a prior survey of project professionals’ LOC affect on perceived role conflict and job satisfaction. A subsequent case-study provides a deeper understanding of the issues that individuals face when working in this extended environment, which could not be fully explored by the scope of the original quantitative study. Findings from both the survey and case study demonstrate LOC attitudinal differences. The survey results signal the important relationship between workers’ locus of control and job satisfaction; in particular, contrary to conventional team findings, those with a sense of internal control appear directly affected by role conflict. In the case study, this may be explained by internals’ willingness to take ownership of their surroundings and actions. This study’s integrative approach highlights the importance of understanding workers’ locus of control orientations and their implications for distributed team member motivation and development. Ó 2006 Elsevier Ltd and IPMA. All rights reserved. Keywords: Distributed working; Virtual team; Distributed team; Locus of control; Role conflict; Job satisfaction 1. Introduction Rapid advances in information and communication technologies (ICT) means that we now live in an electronically-connected world that is seldom ‘asleep’. Organisations and their staff work to satisfy rising consumer and shareholder demands. The 24/7 workplace is increasing, demonstrated by cross-boundary relationships, emphasis on immediacy in everyday transactions, and growing adoption of at-distance cooperative team working. The need for flexibility, supported by a technology platform that allows * Tel.: +44 1483 689347; fax: +44 1483 689511. E-mail address: [email protected]. 0263-7863/$30.00 Ó 2006 Elsevier Ltd and IPMA. All rights reserved. doi:10.1016/j.ijproman.2006.01.003 asynchronous and synchronous communication, has resulted in the growing use of a combination of multi-location and multi-cultural project teams – now commonly known as ‘virtual’ or ‘distributed’ or ‘dispersed’ teams [1,2]. For the purpose of this paper, the terms distributed teams and virtual teams are used interchangeably. Current research on distributed working is centred mainly on systems infrastructure, management control and performance measurement [3,4]. The operational focus is driven largely by the adoption speed and spread of telecommuting across industries [5,6]. As technology-enabled remote working matures to include virtual teamworking, there is a need to look beyond operational issues to understand the psychological and emotional drivers underpinning the way individual workers make sense of their L. Lee-Kelley / International Journal of Project Management 24 (2006) 234–243 environment. Kraut et al. [7] have already cautioned that while electronic networks have made it easier to work virtually, personal linkages remain crucial to mitigate any perceived negative consequences. In addition, Ahuja [8] and Rennecker [9] both call for research to be focused on the place of community within the virtual team and to examine social relations between individual. Hence studying the individual will give us a valuable insight into the personal motivations of virtual team workers and should lead to more effective ways of selecting, motivating and managing them. Katzenbach and Smith [10] propose that virtual teams require different rules of engagement than conventional teams. This suggests that the virtual workers appreciate the reasons for them and will interpret and react to the rules and protocols in a uniform way. Human variability is a recognised phenomenon in psychology and social-psychology, and people have a tendency to react or perceive the world around them differently. How individuals see the world (whether enabling or constraining) will influence their general expectations of whether their own actions would produce the desired outcomes [11,12]. In addition, human behaviour is guided by ‘‘reinforcements’’ (rewards and punishments) and through reinforcements individuals come to attribute their actions or others to cause events that happen to themselves and others. Therefore, in order to understand how team members react and behave when working in distributed project teams, this paper uses the locus of control (LOC) construct to explain the different attribution tendencies underlining virtual workers’ attitudes. Researchers have recently begun to apply locus of control (LOC) in the technology-related domains. Some examples are: Potosky and Bobko [13] studied the relationship between LOC and computer experience; Bellman [14] used the same construct to examine home technology adoption and Sohn and Lekenby [15] investigated the perceived interactivity of the Web. Despite its wide-ranging application in a variety of social and business contexts, using LOC to understand professionals engaged in project-work is unknown. Since job satisfaction is a perception of the stress outcomes from the individual’s appraisal of his/her immediate environment, LOC has been theorised as a good stress moderator [15]. With its basis in reinforcement or learning theory, LOC directly affects behaviour and performance, and offers a robust theoretical framework to research and inform business practice on distributed teamworking. The remainder of the paper is organised as follows: The next section defines the virtual or distributed team and highlights some of the issues facing these teams and their possible impact on worker satisfaction. A review of the LOC literature is then conducted to derive this paper’s theory on at-distance team working. This is followed by a description of the hypotheses and research methodology. The findings of a prior exploratory survey conducted to locate the theory in practice are reported. The survey 235 results are then compared with the findings of a case-study using in-depth interviews to gain further insights. The paper concludes with a discussion of the project’s relevance to management practice and suggests a research agenda for further exploration. 2. Distributed teams Early writers on virtual or distributed teams indicate a competency-based model whereby groups of individuals and/or organisations envisage economic benefits in using modern computing and communication technologies to ‘team’ together to work on a set of common goals and deliverables, but without having to be sited in close-proximity of one another [2,17–19]. Lee-Kelley [20] provides a comprehensive description of what is a virtual team: it is an effective form for delivering large strategic, operational or commercial projects involving various concurrent and sequential activities by team members across various geographical localities. It is likely to be organised in some variant of a matrix structure and may or may not be located as part of the main organisational hierarchy. Often, it is simply an ‘overlay’ onto the existing organisational structure. For larger undertakings, team membership is unlikely to be centrally situated or belonging to a single company, especially when partners are involved. The team itself morphs in size and composition, where a handful of core members would form the nucleus of the team with other members joining or leaving as their contribution or expertise are required or completed. From the above, ‘virtual’ can refer to both physical proximity and the employment or contractual status of participating agents. This does not mean that virtual projects are by necessity, global projects; many are distributed across organisational borders as companies move towards strategic partnering and outsourcing, joint venturing and customer-supplier relationships [10]. For individuals, the practice of working for a variety of (usually non-competing) firms offering their knowledge and skills on projects is increasingly common [20,21]. Often, the individual is employed by one company (the outsource agent or broker) but in fact applies his expertise on a day-to-day basis for the benefit of another organisation (the client); that is, serving as a virtual member of the client’s project organisation. This model of virtual organisation is a familiar feature in IT and clinical research projects, and in academic R&D and defencerelated projects. As team members continue to face many challenges similar to those of their traditional face-to-face colleagues, they also encounter specific issues such as team composition variability, reliance on distance technology, multiple relationships and the logistics of boundary-spanning interactions. The use of outside contractors and partners adds complexity to the virtual team process. The literature on virtual teams identifies self-management (or self-governance) as a key empowerment component of virtual teams. This implies that a team member 236 L. Lee-Kelley / International Journal of Project Management 24 (2006) 234–243 whose prior experience has reinforced his/her feeling of being ‘in-control’ of his work situation and who has reasonable expectations of a positive outcome, is likely to react favourably to his/her environment. Such an individual is more likely to perceive the organisational structure, communication and prescribed work behaviour policies as enabling rather than constraining. However, virtual team researchers have also highlighted barriers such as inappropriate leadership, unequal power and status, extended communication, accountability and responsibility, lack of face-to-face contact, low information or knowledge sharing, poor time management and technology infrastructure and reliability. A further examination of the nature and impact of these difficulties indicates that they have the potential to delay the scheduled outcome of a project. Additionally, delay often requires reprioritising activities and owing to the task-interdependency characteristic of projects, this can lead to confusion, workflow overlap and resource difficulties, compounded as role conflict. The amalgam of these tensions can affect workers’ job satisfaction. Hence, the impact on individual attitudes to role conflict and job satisfaction provides the scope for this study’s first survey. 3. Locus of control Rotter’s [11] internal–external locus of control personality dimension is a result of his earlier work on social learning theory. LOC has been described as a dimension with two opposing differentiates: ‘internal’ LOC individuals interpret reinforcements they receive from their surroundings as contingent upon their own actions, while ‘external’ individuals perceive their actions to be externally determined by say luck, fate or unpredictable factors. The former exhibits a belief in personal control of rewards and outcomes, and the latter attributes consequences to externally imposed factors that are beyond their control. Although LOC research has matured and is applied across a host of industries and work practices [13–15], and Rotter’s [11] original internal–external scale has had its share of alternative scales [22–25], its logic and structure of using a forced-choice format continues to be in use. Arguably, this arbitrary distinction between internal and external control assumes human attitudes to be extreme bipolar, whereas it is just as possible that in reality, a definitive differentiation is hard to ascertain since externallyinclined individuals may exhibit some internal traits and vice-versa. That is, one can be more or less externally or internally inclined; hence the majority of empirical researches on LOC continue to treat control as a continuous variable and using a median split to separate the internals and externals. The hypothesis is that high achievement motivated individuals will be attracted to achievement related activities which allow the attribution of success to ability and effort (i.e., internal or dispositional attributions). These people will persist despite a possible negative outcome, since failure is down to a lack of effort on their part and therefore one must try harder to bring about a successful outcome. Conversely, people who posses low achievement motivation are less attracted to achievement related tasks since failure is more likely to be attributed to the nature of the tasks and luck. This theory is supported by Boone et al.’s [27] study of CEO LOC. In the distributed work environment where direct leader supervision or direction is less apparent, understanding workers’ attitudes to the ups and downs of project life, should help management to provide an appropriate atmosphere for work. Weiner et al. [26] have attempted to explain LOC attributions using a four-factor model of ability, effort, task difficulty and luck. Two of the components (ability and effort) are personal descriptions and the remaining two (task difficulty and luck) depict environmental factors. Together they continue to represent the internal–external LOC dimension. Weiner et al. also highlighted the importance of stability vs. instability since ability and task difficulty are seen as stable, enduring characteristics, whereas effort and luck are relatively variable. This approach is pertinent to the distributed team since its temporal structure assumption means that stability is at a premium and the four factors may be used to explain the interrelatedness of tasks and roles. The attribution conditions under which the outcome of an activity or event is related to each of Weiner et al’s four LOC components are significant in the distributed context: (a) perceived ability at a particular task is a function of the success of prior experience. Hence, if an individual had enjoyed success previously working in a similar environment, he/she will be more agreeable to the possibility of future virtual work; and (b) task difficulty is inferred from the performance of others at the same or similar task. A similar task that is easily completed by others will not be perceived as difficult. If others who had attempted the task before had failed, then personal failure is attributed externally to task difficulty. However if the outcome is positive despite the perceived norm of a difficult task, then eventual success is attributed to personal characteristics and will engender evaluative judgement of individual ability. Luck is inferred from a variable pattern of past outcomes while effort is dependent upon cues such as the level of persistence required for a task. Success leads to an attribution of high effort and failure an attribution of insufficient or low effort. The virtual team structure and member composition as explained earlier in this paper suggests that many if not all of the distributed workers will have to undertake roles and activities that include interactions with team colleagues outside of their own department, division or organisation. Therefore, apart from lack of physical proximity and stability issues, virtual team members have also to manage multiple relationships with internal and external stakeholders when performing their task-based roles. Do internal– external LOC individuals react differently to the need to cope with numerous and often, fleeting relationships? Will L. Lee-Kelley / International Journal of Project Management 24 (2006) 234–243 they relate success or failure to their personal ability and effort or to the external properties of interpersonal task dependency and structural restrictions, or simply to providence? The key question is the extent to which virtual team members’ LOC orientations influence their attitudes to atdistance working and their performance. Many distributed worker are high-skilled professionals who may not be directly employed by the lead organisation [19] and are much harder to manage and sanction in the event of below-par performance. This underlines the need to identify internal LOC people who are high achieving and to provide appropriate personal development to help those with external LOC inclinations to moderate their attitudes. Phares’ [28] study adds further weight to the need for high-achieving personnel. It found that external LOC subjects (those who attribute outcome to external factors) are less realistic than internal subjects in their handling of success and failure and are less capable of coping with their environment. Presumably, as opposed to an individual who enjoys a high sense of internal-LOC and has a positive approach to life in general and his immediate surroundings in particular, someone who does not attribute reinforcements he/she receives to his/her own behaviour is unlikely to be flexible enough nor be willing to modify his/her behaviour to make best use of the particular environment he/she is in. Failure is also linked closely to a high expectancy of punishment. This can lead to stress and anxiety and, at its worst, to a clinical concept called ‘learned helplessness’ [29], thus further reinforcing the perception of a lack of control and that problems at work (and possibly at home) are insoluble. Extreme LOC attributions can become self-fulfilling prophecies and can have disastrous effects on individual and ultimately team performance. 4. Role conflict and job satisfaction The Magerison McCann Types of Work Model identifies eight main types of work in any job role and relates them to individual work preferences, and the actual time spent on each activity type [30]. By showing any mismatch between individual preference and actual work performed, it can act as an indicator of possible role stress. The linking skills index may then be used to encourage companies and teams to better match skills and highlights areas for personal development [30]. However, the self-report basis of the model does mean that it can only ever be indicative, and further qualitative probes are required to examine the reasons for discrepancy. Economic pressures often result in workers being asked to perform roles and tasks that they may not be emotionally suitable or technically equipped to do. Few people are comfortable with door-to-door ‘cold-call’ selling, and a manager with a business degree is more likely to be challenged by the technical aspects of a complex engineering project. In the virtual team context, some individuals may prefer human company to working alone, and others 237 are unhappy having to have to work through the electronic media, particularly for query and dispute resolutions. The remoteness of distributed working also reduces the opportunity for learning and ideas sharing. ‘Sitting-by-Nelly’ is no longer practicable. In a world where individual competitive advantage is dependent on one’s ability to keep ahead of the latest developments, the reduced opportunity to learn and share can affect job satisfaction. The lack of socialisation and having to work in relative isolation can also increase personal stress and retard group solidarity that is characteristic of integrated face-to-face teams. Thamhain and Wilemon [31] propose seven sources of conflict over the project life cycle; namely, schedules, priorities, manpower resourcing, technology, administrative procedures, personality and cost. Work type preference or skills or psychological mismatches in one or more of the seven elements can quickly cause a project to disintegrate. In terms of the LOC construct, individuals with internal and external LOC tendencies will react differently. Moreover, self-preservation means that workers are unlikely to air their difficulties voluntarily. Instead, coping strategies are adopted with different consequences. Some individuals may simply work harder and longer hours to stay ‘on top’. Others may attempt to explain their lack of performance by ‘passing the buck’ to other stakeholders, or blame events that are outside of their control, or to try and retain information to prevent others completing their tasks ahead of them. Although proffered in a period where the majority of business ventures were conventionally organised, from the discussion on barriers above, it would seem that they are also relevant to the modern context. In virtual teams, members may find themselves operating outside of their zone of comfort as they are called to interact horizontally and vertically with multiple project stakeholders. Many of the interactants will not have a prior knowledge of one another. They are also using technology which may be unfamiliar and the connectivity between the interfaces not entirely compatible or fool-proof. While some practitioners [10] advise that clear rules and procedures can overcome distance working and lack of direct supervision and guidance, this may be true only for projects that are relatively simple and straightforward. The structured properties of rules and procedures can, in fact, prevent intuition and innovation by internal-LOC individuals, and for the external-LOC individuals, they serve as a convenient excuse for sluggish performance. Hence, a good understanding of member–member behavioural influences is as important. Culture and language usage and interpretation differences can aggravate any dispute or misunderstanding. These barriers to collective performance affect workers’ ability to perform, and ultimately, their satisfaction with their jobs. Numerous studies have found workers with internal LOC orientations to perceive their jobs more favourably than externals. Individuals possessing opposite attributes tend to feel challenged and resentful of their ‘predicament’. Hence, a picture is emerging of the type of indi- 238 L. Lee-Kelley / International Journal of Project Management 24 (2006) 234–243 viduals who are likely to be well suited to the demands of distributed working, and whose performance is not going to be affected by shifting membership composition. 5. Hypotheses Hypotheses 1. Relationship between control expectancies and role conflict perception H10: There is no difference between internals and externals’ perceived role conflict. H11: There is a difference between internals and externals’ perceived role conflict. The second hypothesis below seeks to ascertain whether role conflict or locus of control exhibits the stronger relationship with job satisfaction. Hypotheses 2. Relationship between control expectancies, role conflict and job satisfaction H20: There is no difference in the relationships between role conflict and job satisfaction and locus of control and job satisfaction. H21: Role conflict and locus of control exhibit a differing relationship with job satisfaction. 6. Methodology The research is two-staged: a prior survey of professional workers involved in defence projects tested the effects of locus of control on team member perceptions of role conflict and job satisfaction. The quantitative results were then compared with the findings of a case-study of IT professionals using in-depth interviews to elicit a deeper understanding of issues facing individuals that was initially indicated in the first study. 6.1. Survey questionnaire A quantitative research design was used to examine the proposed relationships among the various constructs as indicated by the hypotheses above. An approach was made to one of the largest corporate members of the UK Association for Project Management (APM) operating in the defence sector. Many contracts from the Ministry of Defence (MoD) are large and complex, requiring collaborative efforts from participating consortia. To preserve the confidentiality of the participating organisation and their respondents, references to the organisation will be as ‘Company A’. At the time of the research there were around 1500 project professionals working from over 20 locations throughout the UK on military, technical and other scientific ventures. Logistics issues necessitated a web-based survey with an open invitation to Company A’s entire project population to participate. A web-ques- tionnaire was posted and remained open for a month, resulting in 108 completed and usable questionnaires. From earlier discussion on LOC instrumentation, the decision was to use the Rotter [11] IE Scale as the items are still applicable to the modern setting. This is a generalised scale of 23 items, covering social, political, work and educational beliefs. To contextualise the instrument to the project team domain and specifically, to measure role conflict and job satisfaction, two additional sections were included: Part B collected data on role conflict and job satisfaction and Part C collected data on respondents’ demographics. In Part B, questions 24–31 were questions on role conflict and questions 32–36 were question on job satisfaction. For role conflict, Rizzo et al. [32] 13-item 7-point Likert-style operationalisation was used. Answers ranged from ‘very false’ to ‘very true’. For job satisfaction, part of Hackman and Oldman’s [33] Job Diagnostic Survey measuring overall job satisfaction was adopted. Anchored on a 7-point Likert-scale, the answers ranged from ‘disagree strongly’ to ‘agree strongly’. The mean scores were used in the statistical tests for both sections. 6.2. Case study interviews Company B is the UK services division of one of Europe’s largest solution-providers and is also a corporate member of the APM. Company B’s projects cover the private and public sectors, and often include internal and external collaborative partners. At the time of the study the total of project professionals in the UK IT services division of Company B was around 500. However access to individual project personnel was quite difficult as many of them were sited in different parts of the UK or were travelling; hence only telephone interviews using a semi-structured format were feasible for budget and time reasons. Each of the 12 participants was asked to complete a questionnaire before the interview. Interviews lasted between 30 minutes and an hour. They were transcribed, coded and analysed using QSR NVivo version 2.0. Thematic findings specifically on control, role conflict and job satisfaction are used to elaborate on the results of the first LOC survey. The following section reports interview participants’ reactions to distributed working with reference to: (1) the main characteristics as highlighted in the literature; (2) contextspecific factors or problems as experienced by the person, and (3) the effect of 1 and 2 on their job satisfaction. 7. Results 7.1. Survey Of the 108 complete respondents, 84 (78%) described themselves as working in a project management role with the remainder occupying some line management function within the teams. An independent t-test on locus of control and managerial position did not indicate a significant difference between the two means. In addition, although 239 L. Lee-Kelley / International Journal of Project Management 24 (2006) 234–243 the gender split was male predominant (95 males, 13 females), ANOVA results showed no significant difference between the two groups. Therefore, it was established that all 108 completed questionnaires produced usable data for statistical testing. The LOC scale was measured on a forced choice additive scale with the potential for each respondent to score between 0 and 23; the higher the score, the more external the control orientation. Overall LOC mean was 8.29. The scale was then split at the median (i.e., at 8), all scores registering 8 and below are internals, and 9 and above are externals. The split results reported 56 (51.9%) internals with a mean of 5.50 and 52 (48.1%) externals with a mean of 11.20. This suggested a slight preponderance of internals, and is consistent with literature findings that workers with internal-LOCs are more suited to self-managed teams. Compared with the findings from Hyatt and Prawitt’s [34] study of auditors (mean score = 9.4), professionals in the project setting would appear to be more inclined to internal LOC than the other professions. However, Szilagyi et al. [35] study of medical professionals only registered a mean of 5.89 indicating strong internal control expectancies. This is not so surprising since for medics the day-to-day demand for instant decisions (or diagnoses) require professionals that are confident and not risk averse. Cronbach’s alpha coefficients were calculated before hypothesis testing. Role conflict (0.79) and job satisfaction (0.71) were in line with other studies already identified and LOC (0.70) was consistent with that reported by Rotter. LOC was divided into the two groups of internals and externals. The independent t-test used to ascertain perceived difference in role conflict means did not reveal any significant difference: t(106) = 1.250, p > 0.05. This may be due in part to the role conflict mean score for internals (n = 56, mean 4.493) and externals (n = 52, mean 4.7138) being remarkably similar; although the job satisfaction Table 1 Oneway ANOVA: internals–externals on role conflict and job satisfaction Sum of squares df Mean square F Sig. Role conflict Between groups Within groups Total 1.887 120.149 122.037 1 106 107 1.887 1.133 1.665 0.200 Job satisfaction Between groups Within groups Total 5.493 76.505 81.999 1 106 107 5.493 0.722 7.611 0.007 mean scores for internals (n = 65, mean 5.3571) and externals (n = 52, mean 4.9058) did indicate that the two groups had varying job satisfaction perceptions. This is supported by the results of an ANOVA of control expectancies on role conflict and job satisfaction as detailed in Table 1. Therefore, it seems that LOC is not a major factor in explaining role conflict. However, internals appear more influenced by job satisfaction. Using LOC and role conflict as independent variables, two regression models were derived to determine: (1) how LOC predicts job satisfaction over and above role conflict; and (2) how role conflict might predict job satisfaction over and above LOC. Tables 2a and 2b show the model summaries for the regression exercises. In the first regression assessment, the relationship between LOC and job satisfaction registered as significant (p < 0.01) with R2 = 0.076, and F(1,106) = 8.68, thus implying that LOC can be useful for predicting job satisfaction in this context. However, when the role conflict measure was included with LOC, the prediction did not prove to be significant. The change in R2 over and above LOC = 0.018, F(1,105) = 2.11, (p > 0.05). The result of the second assessment of role conflict over and above LOC was also not significant (p > 0.05) with R2 = 0.027, F(1,106) = 2.99, p = 0.09 (see Table 2b). From the results, role conflict appears to be subordinate to locus of control. Table 2a Model summary linear regression analysis Model R a 1 2 0.275 0.306b a b R2 Adjusted R2 Std. error R2 change F change df1 df2 Sig. F change 0.076 0.094 0.067 0.077 0.8467 0.8423 0.076 0.018 8.676 2.110 1 1 106 105 0.004 0.149 Predictors: (constant), locus of control. Predictors: (constant), locus of control, role conflict. Table 2b Second step linear regression analysis Model R R2 Adjusted R2 Std. error R2 change F change df1 df2 Sig. F change 1 2 0.166a 0.306b 0.027 0.094 0.018 0.077 0.8685 0.8423 0.027 0.066 2.994 7.694 1 1 106 105 0.086 0.007 a b Predictors: (constant), role conflict. Predictors: (constant), role conflict, locus of control. 240 L. Lee-Kelley / International Journal of Project Management 24 (2006) 234–243 Reviewing the variables when they have been standardised, LOC had a b of 0.259 and role conflict a b of 0.136. These values form the predictive equation for job satisfaction from which it was noted that LOC exerted a much stronger influence on job satisfaction than role conflict. That is, job satisfaction is impacted more by the individual’s personality dimension than by conflicts within the project environment. This was confirmed by the significant bivariate correlation which showed an inverse relationship between LOC and job satisfaction, albeit that the magnitude of the correlation is low at r = .273 (p < 0.01). Interestingly, there was a significant inverse correlation between role conflict and job satisfaction for internals (r = .305, p < 0.05). Although the literature points to internals as less susceptible to their environment, this finding indicates that in the distributed team, internals are affected by perceived role difficulties and this feeling has a negative impact on their job satisfaction. Although at the aggregate level role conflict appears as a non-significant relationship with job satisfaction, analysis at the dimensional levels (i.e., internal versus external) did reveal a significant relationship between role conflict and job satisfaction for internal-LOC individuals. This result runs counter to the theory purporting that internals’ believe that they are masters of events and difficulties may be overcome by greater effort. In the distributed mode, it is conceivable that task and relationship interdependencies can create role conflict that even with more effort cannot always guarantee a successful outcome. In the light of the above statistical examinations, the difference (albeit it marginal) between role conflict perceptions by internals and externals means a failure to accept the hypothesis H10 that, ‘‘there is no difference in role conflict perception’’. However, the relatively weak results will require further investigation. Similarly, in the case of the amount of contribution to the variance in job satisfaction where it was hypothesised that both role conflict and LOC exert similar impact on job satisfaction, correlation and regression results showed that job satisfaction is impacted more by team members’ control attributes than by conflicts within their work environment. The notable exception is internals’ attitudes to role conflict. Finally, posteriori correlation tests comparing LOC and project professionals’ demographics revealed a relationship between professional qualifications and LOC orientation t(103) = 2.119 (p < 0.05). The problem of ascertaining a project personnel’s professional competence since many merely stumbled into the role had necessitated the use of the APM’s membership list. In the present survey, 47% of the sample had a formal professional standing in project management; which would appear to be considerably higher than the 26% average given by the Association. This higher proportion of professionally accredited, internalLOC project personnel may be a factor of project complexity and virtual teaming characteristic of defence-contracts, as either would demand competence and a ‘can-do’ approach to problem-solving. 7.2. Interviews Initial interviews conducted with three senior managers revealed frequent project over-runs owing to a lack of understanding of clients’ systems as the underlying reason for underestimation of required infrastructure and operational budget and subsequent poor decision-making. This resulted in the need to invest in additional functions to plug the gaps of the core system late on in the projects. The concern that delivery delays and cost escalations would jeopardise long-term relationships with clients was evident, and explains their agreement to participate in the study. The interview participants’ mean LOC orientation (8.44) compared favourably with the original survey sample internal–external mean (8.29). Weiner et al.’s [26] four-factor model was then used to locate some ‘typical’ expectancy responses. Inference to a participant’s LOC orientation is from the consistency of his/her responses. Responses are ordered under the internal–external dimensions according to whether their answers indicated attributions to personal ability and effort or to the project conditions or situation (or as Levenson [25] have suggested, to powerful others and chance). 8. Discussion The prevailing theme from the interviews is that there is a difference in attitudes towards distributed working. On the one hand, the responses were generally positive, even up-beat: ‘I’m not worried’, ‘I like being my own boss’, ‘I welcome it. . . working this way can be fun’. On the other, while accepting that working with others who are not in close proximity and who may not be part of the same organisation has become a business necessity, there is a distinct feeling of discomfort and unease: ‘Sometimes this can result in poor relationship with fellow team members’, ‘. . .if you don’t eyeball someone, you really don’t have an appreciation of their workload’, ‘It can be lonely’, ‘. . .emails are not the same as speech’, ‘How can we create commitment this way?’. This difference was not as evident in the survey since the split between internals (LOC = 5.50, 51.9%) and externals (LOC = 11.20, 48.1%) were close. Similarly, there was only a marginal difference in role conflict mean scores between the internals (4.493) and externals (4.7318). Regression analysis of role conflict and job satisfaction shows no significant relationship, suggesting that team members may view role conflict as an integral part of the job as was reflected in the case-study: ‘Sure, there are always cock-ups and technical issues’, ‘. . ..yesterday I overhead a heated exchange’, and referring to the difficulty of working with third party individuals – ‘It’s particularly hard with contractors and partners – they have their own priorities and agendas’. Of note is the dominant attribution by externals that role conflict issues are caused by ‘others’. This was demonstrated by their persistent use of the third party pronoun ‘they’. Those exhibiting L. Lee-Kelley / International Journal of Project Management 24 (2006) 234–243 internal personalities tended to use the first person collective pronoun ‘we’ in describing their perceptions of role conflict and its possible impact. In the survey, role conflict was measured on a variety of factors, primarily in relation to coping with: (a) numerous tasks and often competing requests but without the necessarily resources or skills to complete them; (b) working with two or more groups of people who operate quite differently; (c) having to work to different expectations; and (d) working on trivial or superfluous things. From the interviews, effects of Thamhain and Wilemon’s [31] seven sources of conflict over the project life cycle were clear from the references to work schedules and priority difficulties, time and resources concerns, reliance on technology, personality and communications problems. Certainly, external participants’ attribute their performances to these characteristics of distributed working: ‘there’s no opportunity to acquire the project-specific skills’, ‘when they [ICT] don’t work, the project grinds to a halt and this can very quickly frustrate [the] workers’, ‘They have me working on 3 projects . . . I never know whether I am coming or going’, ‘It’s having to pander to everyone else’s priorities’, ‘ . . . .they have their own priorities and agendas and affect my ability to do my job’. While acknowledging project difficulties had an effect on their performance, these externals did not see the need to take personal responsibility – ‘if they are not pulling their weight, why should I?’ Participants whose responses indicated their internalLOC disposition, also felt that conflict could be present but the difference is that they were focused on overcoming the perceived difficulties: ‘the key is to have well-organised face-to-face project initiation meetings’, ‘Short lead-in to projects. . . can be managed’, ‘we can find our way around pretty easily’, ‘its down to people to reach out and use their brains’, ‘individuals should take the time to build trust and understanding’. In the survey, when the effect of LOC was controlled, an inverse relationship emerged between role conflict and job satisfaction for internals, suggesting that the greater the role conflict perception, the lower the job satisfaction. Surprisingly, role conflict did not appear to affect externals’ job satisfaction. In this respect partial support is provided for the studies of O’Driscoll and Beehr [36], Newton and Keenan [37] and Spector [38]. Apart from acknowledging its presence the case-study is less clear on internals’ job-satisfaction reactions in a high conflict situation. Prior experience may be a mitigating factor: ‘I have been working in teams made of internal and external people for a long time’, ‘ . . . we’ve been doing this a long time’, ‘the matrix form has been around for years and we’ve gotten used to it . . . ’. For externals, they may view role conflict as arising from management policies and organisational structures, which are beyond their control: ‘what can we do?’ This apparent apathy could be due to a fear of rejection or being exposed as inadequate: ‘I don’t feel comfortable to ask them . . . in case they think I am stupid’ – not unlike Phares’ [28] concept of the fear of 241 punishment. Therefore, the easier option would be not to internalise any perceived problem and not to accept fault. This denial is a useful self-preservation mechanism and if, as Lefcourt [39] suggests, a team member is able to take on an external dimension when the situation is personally unfavourable, then this may explain the project’s survey finding. The LOC literature suggests that internals are proactive and will work hard to achieve their goals. If that is the case, given the balance of internals and externals in the survey sample, is the internal project member being frustrated in his/her attempts to resolve conflict issues by the introverted, external attribution tendencies of his colleagues? Another possible explanation could be that despite the best of intentions, internals’ abilities or skills were somehow wanting and prevented a successful outcome. In the case-study the theme of self-determination was consistent throughout the internals’ responses, regardless of whether they possessed the requisite capabilities, whereas the call for training and development was very clear by the externals. Internals appeared to be optimistic about their own ability and others’ professionalism and best intentions: ‘We are professionals. I am keen to do a good job [so as] not to let my fellow professionals down [and] I expect the same of them’. This positive perspective is encouraging for management. However, as the prior survey had indicated that internals can be overwhelmed by role conflict, it is advisable for management to ensure a match between aspiration and competency. Although the survey and the case-study did not ask specifically about the relationship between job satisfaction and intention to stay, interview participants appeared relatively at ease in their jobs and neither internals nor externals had expressed a dissatisfaction that could cause them to entertain the possibility of leaving. For internals, distributed team working definitely has its compensations in terms of flexibility and self-management: ‘Being a working mum, I find working from home a great bonus’, ‘I like being my own boss . . . .as long as I deliver, I can do what I want’, ‘I feel that this is right, we are professionals after all’. Externals found this freedom and the need to be self-organised a bit daunting: ‘Call me old fashioned, I like working at the office’ ; some might even attempt to turn this into a control issue: ‘It’s important to see what others are doing’, ‘I feel sometimes that all the modern gizmos are there to keep us working 24/7’. Of the post hoc tests that were performed, the one which yielded a significant relationship was that between team members and their professional qualifications. In the survey, more internal respondents held a professional qualification than externals. In the case study, 7 of the 12 participants (58.3%) were individual members of the professional association, APM. As with the survey sample the case-study’s higher proportion than overall average (26%) may be due to IT services projects requiring a higher level of competence and know-how. 242 L. Lee-Kelley / International Journal of Project Management 24 (2006) 234–243 9. Managerial implications From the findings of the survey and case-study, the moderating role of LOC on distributed team worker outlook is demonstrated and should not be overlooked. Although role conflict appears to be an accepted by-product for distributed projects, it does exert pressure on internal-LOC professionals; the reasons of which are still not entirely clear, although previous experience is a possible influencing factor. The casestudy lends support to the sources of conflict identified by Thamhain and Wilemon [31] and managers accountable for project success would be well advised not to rely on team members’ own coping strategies. Closer examination of the survey and case-study results in relation to the existing literature suggest that internals’ enthusiastic approach to their surroundings and problems could be thwarted by their over-positive judgement of their own abilities to bring about improvements. Therefore, while a certain level of self-management is practical and necessary when managing at a distance, it would be prudent to have regular and structured project status and performance reviews. The fairly even split between internals and externals could have made it difficult for internals to bring about the changes that they had envisaged due to the reticence of the latter. This can lead to frustration and job dissatisfaction and ultimately their departure. For management, the findings of this project highlight the importance of team member selection and matching for successful cooperation. The case study has shown that prior experience of technology and/or working in a similar virtual context reinforces control perceptions in the present team and ‘will help people start working and sharing more quickly’. This suggests that it would be useful to have some members of the team who had been contributors to other successful virtual projects. For those without the advantage of previous experience, mentoring and coaching could help people transition from conventional to virtual working. The premise of social learning theory is that a personality dimension such as LOC is relatively but not absolutely stable, in which case it is possible to moderate LOC orientation through a planned personal development programme. Such a programme should enable individuals to recognise their own strengths and weaknesses and to give them the tools to overcome their anxiety of difficult or unknown situations. For those who are uncomfortable or intimated by at-distance cooperation and communicating through artificial media, a strategy for adaptation is required. That is, for externals ‘who like some company’ or ‘prefer working with people I know’, team-building assistance should be given to increase their appreciation for their own and others’ professionalism, encourage their willingness to be proactive, and cultivate a greater sense of community. 10. Limitations and future research The survey was of a cross-sectional self-report design relying on honesty of individuals. The case study’s inter- view participants, although drawn from the same corporate membership listing of the APM, were not part of the original survey. Accordingly, the project’s findings and conclusions can only be indicative and may not be generalisable across other project personnel or other project-based contexts. However, logical extrapolation is feasible, given that both Company A and Company B projects share many common aspects: they tend to be leading-edge and complex undertakings that are organised in a matrix variant involving external and internal partners, subject to the constraints of time, budget and quality and require strong project management competence. In addition, the locus of control questions in the survey instrument were based entirely on Rotter’s [11] expectancy measurement and the questions for role conflict and job satisfaction were extracts from other long-established instruments, as demonstrated by their robust internal consistencies. The lack of strong significant correlations across the groups also suggests the absence of any pervasive underlying correlation-inflation bias. Nevertheless, caution should be exercised in interpreting these data since the narrow quantitative difference is not supported by the wide differentials of view in the case study. It is suggested that future replications in different industry settings will help determine the external validity of this study. Finally, while in agreement with Rotter that new LOC scales must have robust theoretical underpinning, it could be argued that the time has come for a LOC instrument that is specific to the modern networked society, characterised by extensive cross-border collaborations. Acknowledgements This paper would not have been possible without the input and assistance of A. Gilbert who had collected the survey data and K.L. Loong whose access to Company B provided the data necessary for the case study. I also thank the two anonymous reviewers who had generously given their time and whose helpful comments and additional references have shaped the final paper. References [1] Grenier R, Metes G. Going virtual, moving your organization into the 21st century. Englewood Cliffs (NJ): Prentice-Hall; 1995. [2] Lipnack J, Stamps J. Virtual teams: reaching across space, time and organization with technology. New York: Wiley; 1997. [3] Raghuram S, Garud R, Wiesenfeld BM. Telework: managing distances in a connected world. J Strategy Business 1998;10:7–9. [4] Wiesenfeld BM, Raghuram S, Garud R. Communication patterns as determinants or organizational identification in a virtual organization. Org Sci 1999;10:777–90. [5] DeSanctis G, Monge P. Communication processes for virtual organizations. Org Sci 1999;10:693–703. [6] Duxbury L, Haines GH. Predicting alternative work arrangements from salient attitudes: a study of decision makers in the public sector. J Bus Res 1991:23. [7] Kraut R, Chan A, Burtler B, Hong A. Coordination and virtualization: the role of electronic networks and personal relationships. J Comput Media Commun 1998;3(4). L. Lee-Kelley / International Journal of Project Management 24 (2006) 234–243 [8] Ahuja G. Collaboration networks, structural holes, and innovation: a longitudinal study. Admin Sci Quart 2000;45:425–55. [9] Rennecker JA. Reconciling virtual team theory and practice: embedded meanings, structures and networking contexts. In: Paper presented at the 2002 academy of management annual meeting, August 9–14, Denver, CO; 2002. [10] Katzenbach JR, Smith DK. The discipline of virtual teams. http:// www.leadertoleader.org/leaderbooks/L2L/fall2001/katzenbach.html; 2001 [accessed 08.01.06]. [11] Rotter JB. Generalized expectancies for internal versus external control of reinforcement. Psychol Monogr 1966;80(1):1–28. [12] Lefcourt HM. Internal versus external control of reinforcement: a review. Psychological Bulletin 1966;64(4):206–20. [13] Potosky D, Bobko P. A model for predicting computer experience from attitudes toward computers. J Bus Psychol 2001;15(3):391–404. [14] Bellman LM. Using personality traits to predict communication technology use in the home. Doctoral dissertation, Pace University, abstract in Dissertation Abstracts International 1998;58(11-B):6268. [15] Sohn D, Lakenby JD. March. Locus of control and interactive advertising. In: Paper presented at the annual conference of the American academy of advertising, Salt Lake City, Utah; 2001. [17] Cohen SG, Bailey DR. What makes teams work: group effectiveness research from the shop floor to the executive suite. J Manage 1997;23:239–90. [18] Jarvenpaa SL, Knoll K, Leidner DE. Is anybody out there? Antecedents of trust in global virtual teams. J Manage Inform Syst 1998;14(4):29–64. [19] Townsend AM, DeMarie SM, Hendrickson AR. Virtual teams: technology and the workplace of the future. Acad Manage Execu 1998;12(3):17–29. [20] Lee-Kelley L. Situational leadership: managing the virtual project team. J Manage Develop 2002;21(6):461–76. [21] Oates D. Outsourcing and the virtual organization. London: Century Ltd.; 1998. [22] Chung YY, Ding CG. Development of the sales locus of control scale. J Occup Org Psychol 2002;75:233–45. [23] Gupchup GV, Wolfgang AP. A modified work locus of control scale: preliminary investigation or reliability and validity in a sample of pharmacists. Psychol Rep 1997;81:640–2. [24] Spector PE. Development of the work locus of control scale. J Occup Psychol 1988;61:335–40. 243 [25] Levenson H. Activism and powerful others: distinctions within the concepts of internal–external control. J Personal Assess 1974;38: 377–83. [26] Weiner B, Frieze I, Kukla A, Reed L, Rest S, Rosenbaum RM. Perceiving the causes of success and failure. In: Jones EE, Kanouse DE, Kelley HH, Nisbett RE, Valins S, Weiner B, editors. Attribution: perceiving the causes of behavior. Morristown: General Learning Press; 1971. [27] Boone C, De Brabander B, Van Witteloostuijn A. CEO locus of control and small firm performance: an integrative framework and empirical test. J Manage Stud 1996;33(5):667–99. [28] Phares EJ. Locus of control in personality. Morristown: General Learning Press; 1976. [29] Seligman MEP. Learned helplessness. Annu Rev Med 1972;23: 407–12. [30] Author/s unknown. Does team building fit the Asia management culture? http://www.orientpacific.com/opcteam2.htm; 2006 [accessed 08.01.06]. [31] Thamhain HJ, Wilemon DL. Conflict management in project life cycles. Sloan Manage Rev, Spring 1975:31–50. [32] Rizzo JR, House R, Litzman S. Role conflict and ambiguity in complex orgnanizations. Admin Sci Quart 1970;15(2):150–64. [33] Hackman JR, Oldman GR. Development of the job diagnostic survey. J Appl Pscych 1975;60:159–70. [34] Hyatt TA, Prawitt DF. Does congruence between audit structure and auditors locus of control affect job performance? Account Rev 2001;76(2):263–74. [35] Szilagyi AD, Sims Jr HP, Keller RT. Role dynamics, locus of control and employee attitudes and behaviour. Acad Manage J 1976;19(2): 259–76. [36] O’Driscoll MP, Beehr TA. Moderating effects of perceived control and need for clarity on the relationship between role stressors and employee affective reactions. J Soc Psychol 2000;140(2):151–9. [37] Newton TJ, Keenan A. The moderating effect of the type A behaviour pattern and locus of control upon the relationship between change in job demands and change in psychological strain. Human Relat 1990;43(2):1229–55. [38] Spector PE. Behavior in organizations as a function of employee’s locus of control. Psychol Bull 1982;91:482–97. [39] Lefcourt HM. Locus of control: current trends in theory and research. 2nd ed. USA: Lawrence Erlbaum Associates; 1982.