Role of the Project Management Ofﬁce in University Research Centres

: University Research Centres (URCs) have become a primary organisational structure in universities for bringing together a critical mass of multidisciplinary research interests that can compete for large, funded research projects and create breakthrough research results. Some of the more successful URCs are now developing specialised project management ofﬁces (PMOs) that can coordinate key activities, from proposal development to project execution, and ensure that research results are disseminated. A key challenge for URCs is to deﬁne what roles, functions, and competencies such a PMO should have. This research identiﬁes a number of key attributes of PMOs that meet the unique challenges of URCs. This paper presents an initial conceptualisation of roles and functions developed from a literature review and that are later tested via a detailed survey among 370 URC participants involved in collaborative R&D projects worldwide. The study suggests that there are three PMO maturity stages: ‘basic’, ‘intermediate’, and ‘advanced’. The resulting conceptualisation highlights six functions for a ‘basic’ PMO stage, an additional ten functions for an ‘intermediate’ PMO stage, and a further ten functions for ‘advanced’ PMO. The research presented provides guidance and decision support to URCs when selecting the role that a PMO should play for achieving tangible and intangible project beneﬁts. Although the study suggests a lengthy list of functions, none of these should be considered in isolation. Most of the functions interact with each other and affect the PMOs’ impact within the URC in various ways. The paper contributes to the transformative and evolutionary nature of PMOs, and illustrates that universities are receptive and even demanding of the need to create an effective PMO to improve the operation of major R&D projects and programs and create greater societal impact by URCs.


Introduction
Markets have never been so competitive and globalised, and for that reason, organisations need to create more innovative and faster response mechanisms to remain competitive and survive [1]. Consequently, organisations are increasingly involved in more multidisciplinary and cross-institutional research, provided namely through University Research Centres (URCs) or institutes [2]. The URC is one of the most attractive external sources of knowledge and technology for industry [3]. They create opportunities for closer relations between universities and industries and for knowledge development and technological advancement [4]. URCs link researchers from different multidisciplinary areas to cope with complex projects [5]. There is no consensus in the literature on the definition of URCs due to their heterogeneity and the wide diversity of objectives and characteristics [6,7].
A URC can be defined as an organisational entity within a university that aims to serve a multidisciplinary research mission [8]. It is usually separated from the departmental between industries and universities, and national or international funds. Secondly, the conceptualisation of PMO roles and functions contributes to PMOs' transformative and evolutionary nature. The paper also shows that researchers at URCs are highly receptive to creating a PMO structure for improving the performance within their URC projects and programs. Finally, this research contributes to practice for supporting the implementation of PMOs in the context of URCs.
The paper begins with a literature review on the context surrounding URCs and PMOs. It then develops an initial conceptualisation of PMO structures to be used later in the study. The research method used for data collection and analysis is then described. Research results based on factor analysis are then presented. Finally, a conceptualisation of the PMO structures for URCs is discussed and concludes with the management implications and underlining pathways for further research.

University Research Centres
URCs are becoming more common to specifically address the increasingly complex nature of scientific problems that require research solutions that span multi-disciplinary and institutional boundaries [20]. URCs typically bring together researchers from several disciplines and ideologies, different institutions (universities, companies, governments), countries and cultures to solve complex scientific and social-scientific problems [37].
In general, universities and faculty members have benefitted greatly from the presence of URCs. On the one hand, URCs can attract new faculty to collaborate by offering resources and additional funding [38] and allowing faculty members to extend their research agendas [39]. On the other hand, URCs improve the quality of university education since they can attract quality graduate students and enhance comprehensive graduate education [16,39]. They also facilitate interdisciplinary research and collaboration between experts [5,7,9]. URCs are a platform for faculty to focus on their research agendas and gain resources not normally available through academic departments [10,40,41].
These alliances between universities and institutions for the creation of URCs have changed the way the organisational structure of URCs is defined and established [42]. Some URCs are housed within an academic department and adhere to the administration of the department. Other centres function as separate entities within the university. They are governed by an external dean or other authority [7,43], which illustrates that the functions of the URC can differ in their organisational structures and hierarchy within the university. URCs are perceived as specific mechanisms by which institutions create organisational bridges that go beyond the limits of cultural and structural differences [44]. URCs help research projects accumulate scientific knowledge and provide support for increased publication productivity [17,45].
While URCs benefit from conducting research projects due to the plurality of their activities, there are also specific challenges associated with managing such research projects that risk project failure [46]. URCs tend to have heterogeneous research projects and can present management challenges compared to the activities developed by traditional academic structures [4]. URCs deal with relevant scientific problems, which require multiple skills and the integration of different disciplinary perspectives [47]. Such challenges require in-depth knowledge, and an integrated application of appropriate management approaches, namely, to deal with the wide variety of researchers involved [15], ensure the alignment among the collaborative partners, and generate the required level of impact from the R&D projects [48].
In the context of benefits and challenges of project management, URCs strive to improve PM systems and create PMO structures that can minimise project failure [23,46,49]. Current research into PM emphasises the importance of the PMO as an organisational unit that acts as a repository of learning and knowledge transfer. A PMO, for example, can ensure that project mistakes will not be repeated. In organisations with PMOs, projects Sustainability 2021, 13, 12284 4 of 17 tend to be more focused and visible, facilitating communication between project teams and top management [23]. Table 1 illustrates an overview of the URC themes covered in literature, highlighting the contribution of these works for each of the eight themes identified: characteristics and types of URCs (T1); role of URCs (T2); recommendations for URCs (T3); URCs funding (T4); motivations to engage within URCs (T5); performance/benefits/impacts of URCs (T6); URCs collaboration arrangements (T7); and management arrangements/governance/PMOs (T8). These works allowed the authors to obtain a better knowledge of the URC context, for which the PMO roles and functions have been designed.  Reference  T1  T2  T3  T4  T5  T6  T7  T8 A PMO structure is a specialised and formal organisational entity that has within its domain several responsibilities related to the management and coordination of projects [50]. These responsibilities may range from providing support functions to direct PM [51]. PMOs started to become more widespread in the mid-1990s, and since then, their number has grown significantly [33,52,53]. The emergence of PMOs is associated with the increasing number and complexity of projects [33]. This significant increase in complexity has generated new challenges for organisations [54].
PMO structures are continually evolving. Fukuyama and Schumpeter's process of creative destruction provides a helpful analogy to describe this phenomenon [55]. Through an economic view of innovation, the authors argue that the capitalist system can be understood as the evolutionary process where firms adapt through creative destruction. Aubry, Hobbs and Thuillier [56] suggest that PMOs adapt to their environment from a contingency perspective, this being a dynamic and intertwined process between strategy and structure [57,58]. There is a bidirectional relationship between the PMO and the organisation in which it operates, i.e., they adapt and evolve together. This process adopts grounded theory as described in Strauss and Corbin [59], where the researcher analyses data to understand a complex social reality through the development of a process [60]. In this approach, the PMO is seen as a temporary state and participates in the development of the future. This approach has been used to explore the PMO as an organisational innovation [23].
Organisations should be well-advised when deciding to implement a PMO. They should not choose based on mistaken or unfounded assumptions about the value of the money they generate or perceived popularity [61]. Although PMO structures are essential in project-based organisations, the underlying logic that leads to their implementation or renewal is not yet fully understood [62]. Noteworthy, there is no single manual on how to establish and run PMOs in organisations successfully. PMOs are different in size (from single-person departments to entities managing hundreds of people), and there can be just one or several PMOs in various places in the organisation, supporting business, operational or strategic activities [63]. The normative presumptions of longevity, the apparent creation of value, and the descriptions of the generic types of PMOs appear to differ from actual practice, offering neither a solid theory nor a pragmatic orientation to managers [61].
The complexity and variety of PMOs have resulted in various interpretations of what the PMO is and what it should be [64]. Despite this, all definitions have a common feature, i.e., the objective of a PMO is to support PM and increase its effectiveness. The effectiveness of a PMO depends on functions being implemented and their adjustment to organisational needs [65][66][67][68]. Therefore, due to each organisation's different structural and contextual dimensions, it is even possible to have different PMOs in structural and functional terms within the same organisation [69].

Typologies of PMOs
The role and functions of a PMO are subject to various configurations established to ensure the transmission of knowledge and the achievement of goals and objectives [22]. The PMO and the organisation must adapt to the necessary changes to help achieve those goals [66]. PMOs are heterogeneous: they vary in size, function and other aspects [70]. Each organisation should consider what role its PMO plays and adapt it to emerging needs [71]. There is a need to ensure that the roles fit within the organisational and strategic context, increase project performance and meet varying expectations [72]. The challenge for organisations is to reconcile internal PM with governance to align with the organisation's strategic objectives [73]. Therefore, the PMO must adapt to changes that help achieve those strategic goals [66]. Ko and Kim [74] argue that strengthening the project strategic alignments with business goals increases the efficiency and performance of PMOs.
Several characteristics of a PMO were found in the literature. In this study, we identified a total of 55 PMO models comprised of 15 typologies. Typologies found in the literature are presented in Table 2. The most common typology, from which most definitions of typologies originate, comes from the Project Management Institute. The Project Management Book of Knowledge (PMBOK) [51] offers three distinct PMOs with differing levels of authority and control over projects: supportive, controlling, and directive. The PMO can operate as a support unit, providing templates, access to good practices and access to information and lessons learned derived from other projects. It can also control projects, by requiring compatibility of tools and models used within the organisation and standardised PM methodologies and tools. Finally, the PMO can have direct responsibility for all projects and is responsible for overall management.  Table 2 highlights that several authors have proposed different models and typologies to classify the services offered by a PMO structure. Each typology presents a set of functions that a PMO structure should perform [88,89]. It is beyond the scope of this paper to describe each typology referenced in Table 2. For further information, the reader is encouraged to review the respective references.
The implementation or reconfiguration of a PMO is a necessary organisational evolution. Usually, this change is part of a broader organisational reconfiguration. It requires a methodology and an interpretive framework that can capture the complexity of organisational change [62]. Aubry et al. [56] argue that to understand a PMO, one must consider its context and evolution. For these reasons, the goal of this research was to study the implementation of a PMO structure within the URC context, for which there is limited understanding.

Initial Conceptualisation
An initial conceptualisation of PMO structures based on available research literature is proposed. Three types of PMOs are defined based on their maturity: (i) 'basic', (ii) 'intermediate' and (iii) 'advanced'. The conceptualisation focuses on the functions attributed to each type of PMO-emphasising the importance of a contingency approach [90] when pursuing a PMO implementation and highlighting the importance of assessing the relevance of the PMO functions for a particular URC context. Table 3 presents the initial conceptualisation of PMO structures for URCs, i.e., the PMO roles and functions, and key sources or references. The order in which each of these functions appears is not random and should be the starting point for each level. What distinguishes one type from another is the growing importance of the PMO in the URC, a more significant number of responsibilities and the positioning of the PMO in the organisational strategy. All the presented functions have been redefined to clarify and distinguish each function from each other.

Research Method
The first phase of this research study consisted of collecting and analysing the relevant literature to identify PMO typologies and functions commonly referenced and their adaptation to the URC context (see Figure 1). This phase resulted in the initial conceptualisation or model. The second phase consisted of conducting a questionnaire survey, which aimed to validate the initial conceptualisation.

Research Method
The first phase of this research study consisted of collecting and analysing the relevant literature to identify PMO typologies and functions commonly referenced and their adaptation to the URC context (see Figure 1). This phase resulted in the initial conceptualisation or model. The second phase consisted of conducting a questionnaire survey, which aimed to validate the initial conceptualisation.

Data Collection
An online questionnaire survey was conducted. Respondents were asked to indicate the utility of the PMO's functions on a 5-point Likert scale, with '5' meaning 'very high' and '1' meaning 'very low'. The questionnaire was divided into three parts, as follows: • Part A-Characteristics of the respondent. Respondents were asked for information about themselves, their experience and work context (e.g., URC type, scientific area of research projects, roles at URC, experience, age, gender). The questionnaire was developed by an online survey software tool, LimeSurvey [97], and disseminated online through the e-mail contacts of a selected random sample. A random sample is a sample where the population is uniform or has similar characteristics, and any element of the population has the same probability to be selected [98]. The studied population included researchers associated with the URCs. The selection was made using the websites of URCs in 20 universities. Among those chosen universities, 510 URCs were selected to disseminate the questionnaire. The questionnaire was sent to 18,909 researchers via e-mail and resulted in a 2% response rating, corresponding to 370 completed valid responses.

Data Analysis
Data were analysed using the SPSS software [99]. The reliability and validity of the data were tested using Cronbach's alpha and factor analysis, respectively. Factor analysis is conducted to explore the relation of the functions within the PMO, i.e., to verify if the questionnaire results led to the aggregation of the functions resultant from the initial conceptualisation. Factor analysis allowed measurable and observable variables to be reduced [100]. It is a primary method to simplify complex data sets [101]. Later, factors were compared with the typology and roles presented in the initial conceptualisation of a PMO.
More than a third of respondents (37%) had an implemented PMO or similar structure. However, most respondents (81%) believe that the establishment of a PMO structure would be helpful for their URC, and they were motivated to collaborate within the roles of the PMO structure, which means that they perceived the value of a PMO in the URC to support their research work. For the remaining respondents, only 10% indicated that there is no need to implement a PMO structure, while 9% have no opinion.

Validity and Reliability Analysis
Before starting the factor analysis, it is necessary to evaluate the factorability of the data collected. For this, it was essential to verify if the correlation of most of the variables is more significant than 0.3. If this happens, it indicates that the data collected is adequate for factor analysis. Then, Bartlett's test of sphericity and a Kaiser-Meyer-Olkin (KMO) measure helped assess the factorability of the data collected. In the KMO test, the KMO index ranges from 0 to 1, and the factorial analysis is assumed appropriate only if KMO is higher than 0.6 for a better indicator of factorability [99][100][101][102]. As for Bartlett's test of sphericity, it should be p < 0.05 to be significant. After all the test results presented favourable values of factor analysis, the Principal Component Analysis (PCA) was performed. To verify the applicability of the data in this analysis and to be able to proceed with factor analysis, it is necessary to confirm that the communalities have values higher than 0.5 [99]. These results were extracted through the SPSS software package. The next step is determining the number of factors needed to represent the data through the 'factor extraction' [102]. Kaiser's test is one of the most commonly used techniques, also known as the eigenvalue rule [99]. Only the 'factors' with an eigenvalue more significant than one should be considered [102]. The 'varimax rotation' method was performed to simplify the interpretation of the results and to perceive which variables are part of each factor. Table 4 summarises the factor analysis steps followed in this research and their results. Table 5 presents the varimax rotation and variance explained.
From the results obtained (Table 5), it can be determined that: • Factor (component) 1: this factor is constituted by ten positively correlated variables: V15, V17, V18, V19, V20, V21, V22, V23, V24 and V25. Therefore, this factor corresponds to the 'advanced' PMO of the initial conceptualisation. Once the final structure of all factors is established, it is necessary to conduct reliability analysis using Cronbach's alpha analysis, presented in Table 6. The results obtained are reliable since Cronbach's alpha values are all higher than 0.7 [103]. Table 4. Factor analysis steps followed and their results.

Factor/Component Cronbach's Alpha
Determine if factor analysis is applicable to data set All items have at least half of more of their correlation > 0.3 All data is suitable for factor analysis.
The data set has the 'excellent' level for factor analysis (If KMO > 0.9).
Bartlett's test of sphericity is significant (p ≈ 0.000) The data is factorable.
All items/variables have communalities above 0.5, except V1, V6, V9 and V21, very near the threshold of 0.5 The data shows factorability.
Determine the number of 'factors' Three 'factors' have an eigenvalue > 1 explaining 62% of the total variance This is a 3-theme construct.

Discussion
After performing factor analysis, the results show only slight differences from the initial conceptualisation that has proven to be very robust. The results do not exclude any of the functions identified in the initial conceptualisation. However, the allocation of the three functions among the types of PMOs varies between 'basic', 'intermediate' and 'advanced' (V6, V8 and V15).
Regarding V6, the highest loading value was 0.483 in Factor 3, identified as the 'basic' PMO. In the initial conceptualisation, this function was assigned to the 'intermediate' PMO. This result makes sense as this function is related to identifying and categorising all the existing projects within the URC. This function should be one of the first initiatives for a PMO to assist the organisation in project classification and prioritisation. Therefore, in the PMO final conceptualisation, this function is moved into the 'basic' PMO.
Regarding V8, the highest loading value was in Factor 3, also related to the 'basic' PMO. In the initial conceptualisation, this function was designed into the 'intermediate' PMO. As Table 5 illustrates, there was a slight difference between the loading values of Factors 2 and 3, respectively, 0.498 and 0.504. Therefore, considering the content of this function, 'Create an information platform for all past and ongoing R&D projects, which requires a high effort to put in place by PMO members, this function was maintained in the 'intermediate' PMO. Additionally, this function also can be seen as an evolution of V6 because, to create a platform with project information, it is necessary to have information related to the characterisation of projects (V6). An 'intermediate' PMO acts as a source of knowledge and a vehicle for enabling knowledge transfer across URC projects from the same or even different categories, facilitating communication between project teams .
Regarding V15, the highest loading value was in Factor 1, related to the 'advanced' PMO. In the initial conceptualisation, this function was designed into the 'intermediate' PMO. From results analysis, it makes sense that this function should be in the 'advanced' PMO since the guarantee of the exploitation of the project results should be performed by a PMO-experienced team close to the organisation's strategic level. So, for that reason, this function is moved into the 'advanced' PMO. Aligning projects with the organisation's strategic objectives is a major challenge for PMOs [73]; however, strengthening the project strategic alignments with business goals improves the efficiency and performance of PMOs [74]. From a sustainability perspective, the exploitation or post-project phase, and perhaps even the project delivery phase, may bring the expected positive impacts and negative impacts. Therefore the PMO might consider introducing sustainable PM practices into the PM processes [104].
Moreover, 31% of the respondents who answered the open question suggested a new function and responsibility: 'Support the submission of funding applications'. Therefore, this responsibility was placed in the 'intermediate' PMO together with the function 'Support the development of technical and financial reports' (V16). The two functions complement each other since accepted funding applications often require the completion of technical, financial reports later. One of the main motivations for faculty members to join URCs is the access to additional funding [38] to broaden their research agendas; however, these additional resources come also with additional administrative work, such as funding applications and later with regular technical and financial reports, where the PMO can provide important support. Different answers to the open question pointed to functions, such as 'Negotiate the contract with funders and industry' or 'Create a database of potential funding agencies/calls for funding by research area', only referred to by one respondent not considered in the final conceptualisation.
Therefore, a final conceptualisation of PMO structures for URCs with a total of twentysix functions is proposed in Figure 2. These are divided into the three PMO typologies and include six for the 'basic' PMO, a further ten for the 'intermediate' PMO, and an additional ten for the 'advanced' PMO. The conceptualisation presents an evolution logic, which means the PMO role might evolve along the time from 'basic' PMO to an 'advanced' PMO.
Besides the gradual implementation of the PMO's functions [105], a PMO structure also depends on some other vital factors: 1. PMO functions' alignment with the URC's culture [62][63][64][65][66][67][68][69][70][71][72][73][74][75][76][77][78][79] and strategic direction [46]. For example, in the context of a URC, the title of Project Manager is not commonly used since Professors/Researchers associated with URCs with a scientific background typically do not see themselves as managers. URCs might consider naming them Project Leaders. 2. URCs' top management needs to support the PMO implementation [106] and recognise the value of formal and standardised PM practices set by the PMO [52,83,105]. 3. Choice of the right leader of the PMO structure and the right PMO team . 4. Definition of clear communication channels between the PMO structure and the URC top management and project teams [79][80][81][82][83]. The development of communication skills at the individual level is one of the most critical barriers to knowledge transfer [93].  Besides the gradual implementation of the PMO's functions [105], a PMO structure also depends on some other vital factors: 1.

2.
URCs' top management needs to support the PMO implementation [106] and recognise the value of formal and standardised PM practices set by the PMO [52,83,105].

3.
Choice of the right leader of the PMO structure and the right PMO team .

4.
Definition of clear communication channels between the PMO structure and the URC top management and project teams [79][80][81][82][83]. The development of communication skills at the individual level is one of the most critical barriers to knowledge transfer [93].

Conclusions
This research has focused firstly on the role and functions of PMOs within University Research Centres (URCs) based primarily on previously published literature. Secondly, it reviewed the functions adapted to the URC context, based on the researchers' judgment and a PMO initial conceptualisation. It then deployed a web-based questionnaire survey where 370 respondents offered their views and experience on the perceived usefulness of the functions identified in the initial conceptualisation. Finally, through factor analysis, this research validated much of the PMO proposed conceptualisation. Twenty-six key functions were presented in URCs, divided by three PMO typologies: 'basic', 'intermediate' and 'advanced' (see Figure 2).
The questionnaire results also show that respondents perceived the value of a PMO and that URC members are receptive to creating a PMO in their own context. As Artto et al. [50] discussed, the PMO can have an integrative role in the front end of innovation and contribute to the overall organisational performance discussed by Aubry and Hobbs [107]. More recently, Sergeeva and Ali [34] demonstrated that PMOs play an essential role in integrating innovation, exploration and exploitation throughout the project life cycle. Project managers, commonly named Principal Investigators in URC organisations, are excellent researchers but less skilled or interested in PM [35]. Therefore, PMO structures can play a critical role in the URC context.
The research reported in this paper develops new knowledge in the domain of PMOs in the context of URCs, for which there is currently limited understanding. This study identifies the primary roles and functions of a PMO to support URCs. Moreover, it provides empirical evidence on the evolutionary perspective of the implementation of PMOs within organisations.

Management Implications
This research study provides guidance and decision support to URCs when selecting the role that a PMO should play for achieving tangible and intangible benefits of their projects and programs. It provides important managerial insights by indicating some of the competencies, usually outsourced by URCs to management consultancies, that might be performed instead by an internal structure such as a PMO. By contrast, Martins and Martins [108] explore the mechanisms that influence decisions regarding outsourcing competencies in the operation of PMOs.
PMO structures support the implementation of PM practices that would help to maximise project and program benefits, and to increase the transparency of information transfer among research partners [65]. The PMO promotes trust, which in turn contributes to reducing the effect of different and sometimes competing expected benefits from research partners [109]. The PMO plays a significant role in embedding PM practices, particularly where there is a high interdependence between partners [110]. Moreover, the PMO acts as a repository of learning and a vehicle for enabling knowledge transfer across URC projects, thereby achieving greater project synergies .
PMOs have an important contribution to project start-up, execution, and societal impact by establishing a pseudo-independent unit focused on strategy, ideation, project management, project portfolio management, and enhancing societal impact. Establishing a PMO is not without costs, but the benefits are clear both from this research and literature in the field. The costs for a PMO can be offset by the overhead contribution now common within all major research projects and programs. PMOs can enhance learning around the development and execution of complex programs, improve communications, and perhaps most importantly, ensure that quality and risks are managed and that societal impact is assured in addition to scientific impact.

Limitations and Future Research
The existence of few studies on PMO structures from the perspective of URC organisations was the main difficulty faced by the researchers. Almost all of the literature analysed refers to the implementation of PMOs in an exclusively business context. The relatively low response rate (2%) to the questionnaire survey might be seen as a research limitation; however, this is offset by the high number of respondent participants (370).
Although the study suggests a lengthy list of ordered pertinent functions for URCs, none of these should be considered isolated variables. Most of the functions interact with each other and affect the PMOs' potential impact within the URC. Therefore, further research can be performed by applying the findings of this study towards the understanding of the roles of PMOs and the weight that different URCs place on various functions and how they interact, highlighting the URC contextual variables. It would also be helpful to develop a model to assess the effectiveness and efficiency of PMOs in URCs since the PM value for the particular context of R&D projects [71] has been questioned by several scholars [111,112].
Author Contributions: The four authors collaborated to produce this paper. G.F., as the principal researcher, designed and conducted the research in the field; H.S. supported the collection process of data; A.T. and D.O. provided guidance and support for the fieldwork and the development of the conceptualisation of PMO structures for URCs. The four authors contributed to the analysis and interpretation of the results. All authors have read and agreed to the published version of the manuscript.
Funding: This work has been supported by national funds through the FCT-Fundação para a Ciência e Tecnologia, within the R&D Units Projects UIDB/00285/2020 and UIDB/00319/2020. Institutional Review Board Statement: Not applicable.