The Built & Human Environment Review, Volume 3, Special Issue 1, 2010

The Use of Analytic Hierarchy Process (AHP) in the Analysis of Delay Claims in Construction Projects in the UAE
Perera, N. and Sutrisna, M. 1.nihal@gmail.com and M.Sutrisna@salford.ac.uk School of the Built Environment, University of Salford, Salford, M5 4WT, United Kingdom

Abstract
Central to successful resolution of delay based claims in construction projects is the fair and equitable apportioning of parties liabilities. However, such fairness and equity in apportioning may be fettered by many factors in contemporary practices adopted by both contractors and clients in construction industries including in the United Arab Emirates (UAE). The principal goals and rationale of this research are inspired by the need to identify such potentially major obstructive factors and provide improvements to the problematic situations that may stem from them. One such factor in apportioning liabilities is identified in that there is no universally acceptable delay analysis methodology amongst practitioners. As the research study aims to bring about contractually sound, practical improvements to problematic situations in contemporary practices, an area of such improvements is identified in developing a model for selection of most appropriate delay analysis method under a set of specific circumstances of a project. Such model is expected to provide the practitioners with a more objective and tenable basis for delay analysis which would enhance fair and equitable outcome in apportioning liabilities. In building this model, the Analytic Hierarchy Process (AHP) has been selected as the appropriate decision making tool. Despite its reliance on mathematical manipulation, AHP embraces the use of mixed or multiple methods in its data collection and analysis. The interplay between quantitative and qualitative data and analysis is expected to result in the phenomenon of apportioning liabilities in delay claims resolution being seen from different perspectives illuminated with mixed methods.

Key words: Ontology, Epistemology, Methodology, Decision Making, Analytic Hierarchy Process

Introduction
The experiential knowledge and peer-discussions as to current delay claims resolution in the local industry have indicated that a majority of delay claims in the Unites Arab Emirates (UAE) falls short of adequately discharging the burden of proof for causation. Adding to this are the stakeholders using misconceived or misconstrued legal principles in the apportioning of parties liabilities, particularly where concurrent delays are involved, and the use of inappropriate techniques in delay analysis in apportioning those liabilities. The findings of the Pilot Survey carried out in the research in the UAE have also confirmed the continuing existence of such problematic situations and a strong need to identify them to provide necessary remedies/improvements. This situation has given the inspiration for the current research. Based on a set of research question developed from three types of questions for qualitative research (Maxwell, 1996), namely, Descriptive Question, Interpretive Question and Theoretical Question., the research aims to critically inquire how the apportionment of liabilities in delay claims
29

The Built & Human Environment Review, Volume 3, Special Issue 1, 2010

is practiced in the local context with a view to identifying problematic situations associated with such practices. The outcome of that inquiry is expected to be used towards developing a framework for improvement. Accordingly, the research methodology is developed with the following overall objectives, in response to the key research questions: i. To investigate how the apportioning of liabilities in delay claims is carried out by practitioners, particularly in the local UAE context; ii. To gather evidence of problematic situations stemming from the approaches and methods currently used by practitioners in the UAE; iii. To develop a constructive improvements to deal with such problematic situations in the UAE; and iv. To assess the extent to which the proposed improvements are evident to improve the current approaches and methods in terms of contract documentation, claims administration, and delay analysis methodology. The research design represents the logical sequence that connects the empirical data to research questions and ultimately to its conclusions (Yin, 2003). Accordingly, the initial part of this paper discusses typology of major social paradigms along with Critical Realism perspective as the philosophical position of the research. Following the discussion on ontological and epistemological positions, the focus of the paper shifts to the research methodology designed to satisfy the research aim and objectives. It observes how the ontological and epistemological positions influenced the research methodology. This is followed by a discussion on the initial findings of a Pilot Survey, and the current status of inquiry for data collection and analysis. In the final part of the paper, there will be a brief discussion on Analytic Hierarchy Process (AHP). There, the selection of AHP in the research study, its history, philosophy and performance, and how it is placed within the research is discussed. As a reflective thought on the development of this research strategy, it is noted that the initial philosophical position of the researcher was within the Naturalistic approach and the Constructivist paradigm. However, further studies and personal development of the researcher has led to a position-shift into Critical Realism which has been mainly influenced by the works of scholars such as Archer et al. (1998), Robson (2002), Mingers and Brocklesby (1997), and Mingers (2004).

Establishing Methodological Position

Denzin and Lincoln (1994) define a paradigm as a basic set of beliefs that guides action. Paradigms deal with first principles, or ultimates. Guba and Lincoln (1994) argue that these beliefs are basic in the sense that they must be accepted simply on faith (however well argued) and there is no way to establish their ultimate truthfullness. Each paradigm can be discussed within the context of their three main elements, namely ontology, epistemology and methodology; Ontology raises questions about the nature of reality while epistemology asks how do we know the world; Methodology concerns on how we gain knowledge within the context of the selected nature of reality (Denzin and Lincoln, 1994).

Paradigms, Methodology and Techniques

Guba and Lincoln (1994) conceptualised four basic paradigms as belief systems based on the philosophical assumptions of ontology and epistemology to see the world and how it should be studied and understood. These basic paradigms are identified as Positivism, Post-positivism, Critical Theory et al, and Constructivism. Mingers and Brocklesby (1997) distinguish three main paradigms each of which has been referred to by a variety of

30

The Built & Human Environment Review, Volume 3, Special Issue 1, 2010

names: empirical analytic (positivist, objectivist, functionalist, hard), interpretive (subjectivist, constructivist, soft), and critical (critical systems). A methodology has been defined as developing, either implicitly or explicitly, within a particular paradigm and embodying the philosophical assumptions and principles of the paradigm. On the other hand, a technique is defined as a specific activity that has a clear and well defined purpose within the context of a methodology (for example a statistical analysis). It was suggested that the relation between methodology and technique is that between a what and how.

With regard to Ontological position, Critical Realism finds a unique position in the continuum between realism (objectivist) and relativism (subjectivist); Ontologically, it takes the position that objective world is independent of human beings, existed before humans and would exist with or without them, but at the same time acknowledges that knowing of the objective world cannot be without human perception, experience and abilities. Crotty (1998) argues that it becomes a world of meaning only when meaning-making beings make sense of it; in this line of thought, the existence of a material world without mind is conceivable but meaning of it without a mind is not. Thus, this world can be characterized as objective in the sense that it is independent of the observer (humans), but our observations and descriptions of it are not. According to Bhaskar (1978) Critical Realism has a stratified rather than flat ontology: the empirical, the actual and the real. Mingers and Brocklesby (1997) suggest that from the material world, through process of evolution, linguistically endowed humans have developed, capable of communication and self reflection; and this has led to the social and personal worlds. Ontologically, this personal world is subjective in that it is generated by, and only accessible to, the individual subjects; its a world of our own individual thoughts, emotions, experiences and beliefs; the social world is the one that we share as members of particular social systems. Easton (2009) argues that the difference between Critical Realists and Social Constructionists lies in the acceptance of the possibility of knowing reality in the former case, and its rejection in the latter, who, in general, concentrate on uncovering the constructions that social actors make. Compared with Critical Realism, Positivism, Post-positivism and Critical Theory et al ontologically take the objectivist (realism) position, more or less on one end of the continuum. Positivism is commonly called naive realism by its critics. It identifies with an apprehendable reality which is assumed to exist, driven by immutable natural laws and mechanisms. Knowledge of the way things are is conventionally summarized in the form of time-and-context-free generalizations, some of which take the form of cause-effect laws. The basic posture of the paradigm is argued to be both reductionist and deterministic. In Postpositivism reality is assumed to exist, but to be only imperfectly apprehendable because of basically flawed human intellectual mechanisms and fundamentally intractable nature of phenomena. It preserves the basic assumptions of positivism: ontological realism, objective truth, use of experimental methodology. Critical Theory et al are treated as historical realism and reality is assumed to be apprehendable but over time shaped by a congeries of social, political, cultural, economic, ethnic and gender factors and crystallized into a series of structures that are now (inappropriately ) taken as real, that is natural and immutable. Compared with Critical Realism, on the other end of the continuum is Constructivism. It is based on relativism, that is when realities are apprehendable in the form of multiple, intangible mental constructions, socially and experiential constructions are not more or less true, in any absolute sense, but simply more or less informed and/ or sophisticated. Constructions are alterable, as are their associated realities. Thus, Constructivism differentiates from other three paradigms as it is guided by relativism while the others are by various forms of realism. In Epistemological discourse, Critical Realism finds epistemological relationship to material world as one of observation rather than participation as in social activity or experience as of a personal feeling. As Mingers and Brocklesby (1997) argue, however, in our personal world we do not observe it but experience it. We can aim to express our subjectivity to others and, in turn, appreciate theirs. Then our epistemological relationship to the
31

The Built & Human Environment Review, Volume 3, Special Issue 1, 2010

social world is one of inter-subjectivity; on one hand it is a human construction, but then it goes beyond and preexists any particular individual. Compared with Critical Realism, Positivism, which is an objectivist approach maintaining a dichotomy between the object (the observer) and the subject (the observed), assumes the investigator and the investigated object are to be detached, independent entities, and the investigator to be capable of studying the object without influencing it or being influenced by it. Inquiry takes place as through a one-way mirror. Replicable findings are, in fact, true. In Post-positivism, though such dichotomy between object and subject is largely abandoned, objectivity remains a regulatory ideal; Replicated findings are probably true but always subject to falsification. Constructivism and Critical Theory et al have transactional and subjectivist approach and assume that the investigator and investigated object are to be interactively linked, with the values of the investigator (and of situated others) inevitably influencing the inquiry. Findings are therefore value mediated. In this posture, the traditional distinction between ontology and epistemology is challenged. What can be known is inextricably intertwined with the interaction between a particular investigator and a particular object or group. Accordingly, Constructivism and Critical Theory et al are different from other two paradigms in their approach to the subjectivist - objectivist dichotomy. In Methodological analysis, Critical Realism appears for multi-methods. Mingers and Brocklesby (1997) argue that it is always wise to utilize a variety of paradigms. It was argued that adopting a particular paradigm is like viewing the world through a particular instrument such as a telescope, xray machine or an electron microscope. Although they may be pointing at the same place or thing, each instrument produces a totally different and seemingly incompatible representation. Thus, using only one method, one would get only a limited view of the (problem) situation. Positivism takes an experimental and manipulative approach with carefully controlled (manipulated) conditions to prevent outcomes from being improperly influenced. This methodology thus focuses on verification of hypotheses and uses chiefly quantitative methods. Although this is the general approach, there may also be circumstances where positivists use social statistics for analysis. Post-positivism emphasises on and has mainly invested in critical multiplism (a refurbished version of triangulation) focusing on falsification (rather than verifying) of hypotheses. In Critical Theory et al the transactional nature of inquiry requires a dialogue between the investigator and the subject of inquiry. That dialogue must be dialectical in nature to transform ignorance and misapprehensions (accepting historically mediated structures as immutable) into more informed consciousness. This view aimed at the reconstruction of previously held constructions. As for Constructivism the variable and personal (intra-mental) nature of social constructions suggest that individual constructions can be elicited and refined through interaction between and among investigator and respondents. These varying constructions are interpreted using conventional hermeneutical techniques and are compared and contrasted through a dialectical interchange. The final aim is to distil a construction that is more informed and sophisticated than any of the predecessor construction. Except for that, this also aims at the reconstruction of previously held constructions, similar to Critical Theory paradigm.

Research Issues in Perspective

The above paradigms, as sets of basic beliefs, are not open to proof in any conventional sense and there is no way to elevate one over another on the basis of ultimate, foundational criteria. Guba and Lincoln (1994) argue that any given paradigm represents simply the most informed and sophisticated view that its proponents have been able to devise, given the way they have chosen to respond to the three defining questions appear below: (i)Ontological question Whats the form and nature of reality and, therefore, what is there that can be known about it?

32

The Built & Human Environment Review, Volume 3, Special Issue 1, 2010

(ii)Epistemological question: What is the nature of the relationship between the knower or would be knower and what can be known? (iii)Methodological question: The question here is how can the inquirer (knower) go about finding out whatever he believes can be known? In terms of the said three defining questions, it is required to examine the issues in the research area, from the Critical Realism perspective as the current philosophical position.

Ontological Assumptions
The question here is whats the form and nature of reality and, therefore, what is there that can be known about it? Central to successful resolution of delay based claims is the fair and equitable apportioning of parties liabilities. Therefore, parties primarily desire that the apportioning of liabilities is carried out fairly and equitably. However, whether these aspirations of parties are met or not will largely depend on how the apportioning of liabilities is carried out by the practitioner on both sides. It is already mentioned that the theoretical perspective of the researcher in this research study has shifted from its previous position. The earlier position led to the perception that apportioning liabilities was dominated by subjective meanings or constructed notions ascribed by the practitioners to delaying events and similar phenomena. These meanings and notions were considered as constructed realities rather than objective realities, and dominating the whole of apportioning liabilities process and its outcome. Following the shift into Critical Realist that accepts the existence of a material world without mind is conceivable, the current positioning recognises the tangible entities such as projects, activities, (delaying) events, damages, project duration, terms and conditions of contract which generate the respective rights and obligations of parties, and the like witnessed in the projects as objective reality that exists independent of the practitioners subjective meanings and notions. It is perceived now in the development of research study, the more such constructed notions (subjective) are in harmony with these tangible (objective) realities the higher the chances of successful resolution of delay claims. The current position acknowledges that the practitioners perceptions, experiences, attitudes, judgments and so on embedded in delay claims resolution are not objective phenomena but are meaningful constructed realities only; however, without those tangible entities (objective realities) of projects, activities, (delaying) events, damages, project duration, terms and conditions of contract and so on, these meanings would not have been made, sensed or derived from on their own. Therefore, ontologically, the research ontological position accepts that those tangible entities (objective world) are independent of individuals, but making sense of them cannot be without human perception, experience and abilities. Accordingly, while accepting the objective existence of projects, activities, (delaying) events, damages, project duration, terms and conditions of contract and the like, how the apportionment of liabilities in delay claims is practiced, and what problematic situations are deriving from such practices can be perceived only through the individual practitioners subjective experience, tacit or explicit awareness, and interpretation of what they have implemented in practice. Crotty (1998) suggested that their meaning is not discovered, but constructed. As this obtained perception is built around individual cognition, its existence or being is primarily idealist and relative. Critical Realism ably accommodates the co-existence of the intransitive domain of objective entities like projects, activities, (delaying) events, and the like, as well as their being of objects of transitive domain of practitioners activities, experience, multiple view points and meanings derived from them. Mingers and Brocklesby (1997) point out that Critical Realism depicts the co-existence of intransitive objects of knowledge (entities that exist independent of our experience of them) and transitive objects (our experiences, theories and
33

The Built & Human Environment Review, Volume 3, Special Issue 1, 2010

descriptions that are used in the production of knowledge). Also the research inquiry intends to capture information from multiple approaches and methods of practitioners who may construct meaning in different ways even for the same occurrence (Pluralism); each of these approaches and methods, though they can be divergent and mutually conflicting, may still be meaningful as fair and equitable under different circumstances, providing they are implemented aptly and pertinently. For example, a judgment to use a particular delay analysis technique may be inconsistent with the terms of one particular contract, but it may be consistent with those of another contract. In the current research, these multiple constructed realities are studied holistically. Inquiry into such multiple meanings and viewpoints will inevitably be divergent, comprehensive, variable, and dynamic so that prediction and control are unlikely outcomes although some level of understanding can be achieved. Ontologically, Critical Realism accommodates existence of such multiple subjective meanings within its personal world and social world.

Epistemological Assumptions
The question here is whats the nature of the relationship between the knower (inquirer) and what can be known? Patton (2002) argues that in practice, human interventions are often quite comprehensive, variable, and dynamic, and this creates considerable difficulty for controlled experimental designs that need specifiable, unchanging treatments to relate specifiable predetermined outcomes. The propositions in the current research settings can hardly be converted into precise mathematical formula expressing functional relationships. Human behaviour, unlike that of physical objects, should be understood with reference to the meanings and purpose attached by human actors to their activities. Qualitative data, it is asserted, can provide rich insight into human behavior (Denzin and Lincoln, 1994). In the current research epistemological stance, the subjective meanings ascribed by human experience are considered dominant as they are informed by the practitioners (respondents) and their activities. Those are understood through acceptance of subjective relationship with the human actors. The things that can be known in this research, encompasses the manner that the contemporary practices in the local industry deals with the apportionment of liabilities. This is to be known primarily in order to perceive the problematic situation which is the next thing that can be known. Thereafter, in order to know the necessary improvements required for the existing practices, the views of the practitioners who experience the problematic situation are to be captured by the researcher (knower). This knowledge is being captured through the interaction between the researcher and the respondent practitioners. For that, the knower and the object to be known have to be interactively linked. Thus, the inquiry process cannot be routed through a one-way mirror. Critical Realism accepts the epistemological relationship to deal with the human factor is not of objectivity. Groff (2004) identifies check on relativism is one of the several key epistemic principles of Critical Realism. Archer et al. (1998) suggest that critical realism claims is able to combine and reconcile ontological realism, epistemological relativism and judgmental reality. Similarly, Robson (2002) points out that critical realism seeks to achieve a dtente between the different paradigms of a post-positivist approach within the empirical tradition on one hand, and less thoroughgoing versions found in some constructionist approach on the other.

Methodological Assumptions
The question here is how can the inquirer (knower) go about finding out whatever he believes can be known? Guba and Lincoln (1994) suggest that the answer that can be given to this question is influenced by the answers already given to the first two questions. Accordingly, the methodological question cannot be reduced to a question of methods; methods must be fitted to a predetermined methodology (Denzin and Lincoln, 1994).
34

The Built & Human Environment Review, Volume 3, Special Issue 1, 2010

Burrell and Morgan (1979) suggest that, if the researcher takes an objectivist standpoint to reckon the social world as a hard and objective reality existing outside to the individual cognition (Realism) then his method of inquiry of research would be seeking to explain and predict what happens by searching for regularities and causal relationships between its constituent elements (Positivism); this researcher will treat human beings as product of their environment (Determinism) and his research methodology would be aiming to search for universal laws which explain and govern the reality that is being observed (Nomothetic methodology). Nevertheless, many scholars may not agree to such limitation as to the issue of methodology. Crotty (1998) argues that the distinction between qualitative research and quantitative research occurs at the level of methods and not at the level of epistemology or theoretical perspective. However, as discussed above, the philosophical position of Critical Realism in this research accepts a subjectivist standpoint to reckon the personal and social worlds. It considers the objective world is observed by participants differently with multiple meanings and points of view. These varying subjective meanings constructed by the participants (i.e. practitioners) are to be compared through dialectical interchange in order to distil a consensus construction that is more informed and sophisticated than any of the predecessor construction. This requires a method of inquiry of research that would focus on different issues in different ways to understand from the inside rather than the outside (Anti-positivism) as well as treating human beings as possessing free will to act voluntarily (Voluntarism). Having discussed the philosophical position above for methodology, it requires developing the specific inquiry strategy. A mixed or multi-method is selected and both quantitative and qualitative approaches and methods are discussed here. Tashakkori and Teddlie (1998) saw the emergence of mixed methods as a third paradigm being distinct from the quantitative research mainly embraced within positivist perspective on the one hand, and the qualitative research mainly embraced within the constructivist perspective on the other. Thus, the role of mixed method approach in research inquiry appears to have certain similarity with the role of Critical Realism which has found itself in somehow between positivism and constructivism. A main advantage of employing multiple methods is cited as permitting triangulation. Robson (2002) suggests that multiple methods can also be used in complementary fashion to enhance interpretability. Thus, it may also be acknowledged that the validity of using mixed and multiple methods is not limited to any specific ontological and epistemological perspective. Robson (2002) claims that one important benefit in applying multiple methods is in the reduction of uncertainty. The reality in an applied field like construction delay claims resolution can be considered multiple, complex, constructed and stratified and in many cases needing a higher rigour in the research procedures. The use of mixed-method approach in this research has been influenced by the Critical Realism paradigm that would lead to the adoption of quantitative and qualitative approaches as the method of inquiry. In other words, multi-method approach is accepted in this research as the operationalisation strategy of inquiry inspired by Critical Realism. In this way, the chief aim is to use the methods in a more integrated way and therefore, the different methods are combined in sequential manner. The methods that are to be combined in the current inquiry are mainly pilot and in-depth surveys (questionnaires) and semi-structured interviews (with content analysis of transcripts according to a coding system). Accordingly, this approach is expected to provide not only a basis for triangulation but also a source to look at the same things from different points of view of conceptualising the problem situation.

Current Stage of Inquiry

As for the current stage of inquiry, a pilot survey questionnaire has been conducted to better understand the need of the research and accordingly to galvanised the research aims and objectives. A total of 18 practitioners were selected as potential respondents in this pilot survey. Altogether 12 responses received representing 66% response rate. Amongst these, there were 8 practitioners from consultancies (66%), 1 from contractor
35

The Built & Human Environment Review, Volume 3, Special Issue 1, 2010

organizations (9%) and 3 from Developers (25%). Sector wise, 64% of respondents were from claims management, 27% from QS and commercial management and 9% from contracting background. Of all the respondents 84% were from middle management, 8% from lower management, and 8% from top management level. All 12 respondents consented to further participation of the research. There were seven propositions submitted to the respondents for their opinions. The responses received are as follows: 1. 57% strongly agreed and 37% agreed while 6% were neutral to the proposition that UAE/Dubai construction industry represents the most advanced characteristics and growth rate in the industry of whole of the Middle East region" 2. 65% strongly agreed and 29% agreed while 6% were neutral to the proposition that "Delay claims remain a major source of dispute in UAE/Dubai construction industry, requiring prompt resolution to avoid escalation to major, complex dispute situation". 3. To the proposition that "Delay claims resolution process in UAE/Dubai construction industry is generally protracted and often happens near to or even after the completion of projects" 47% strongly agreed and 53% agreed. 4. 21% strongly agreed and 75% agreed while 4% disagreed to the proposition that Absence of prompt resolution of the delay claims generally escalates those claims to major dispute levels". 5. 21% strongly agreed and 68% agreed while 4% disagreed and 7% were neutral to the proposition that "Between the contractors and the developers/engineers, for delay analysis there is no consensual selection of analysis method that is most appropriate under given circumstances of a project". 6. To the proposition that "In order to make delay claims resolution process in UAE/Dubai industry more efficient, transparent, equitable and fairer, there is a strong need to identify the current practices, their problematic situations and necessary remedies/improvements required to them" 79% strongly agreed and 21% agreed. 7. Again, to the proposition that "Such, remedies and improvements would possibly bring corporate benefits for both developers and contractors, reducing delay situations escalating to dispute levels such as Engineer's Decisions/Arbitration/ Litigation" 79% strongly agreed and 21% agreed. The preliminary findings of this Pilot Survey indicated the need of the current research subject and enabled the articulation of its aims and objectives as mentioned in the earlier part of this paper. This Pilot Survey has been followed by detailed semi-structured interviews with 10 practitioners (so far). Creswell (2007) cites recommendation of Polkinghorne (1989) that researchers interview from 5-25 individuals who have all experienced the phenomenon. Therefore, these detailed interviews involving 10 practitioners who are seriously involved in claims resolution are considered within that recommendation. The interviewees were ranging from delay claims experts to construction lawyers. The interviews have average 90-100 minutes duration. The full transcripts of these interviews are being prepared now for inclusion in the necessary templates with coded themes and sub-themes. Parallel to these interviews, an in-depth questionnaire survey is being conducted. Initially, 520 potential respondents across the industry in UAE/Dubai have been contacted through snowballing technique for this web-linked questionnaire, but over 200 questionnaires returned, undelivered, reducing the potential responses to about 300. Of these, so far, only 57 have responded. Thus, the response rate so far is just around 19%. This seems to be a somewhat within the anticipated response rate considering the vast number of employees laid off in the UAE/Dubai construction industry since the economic crisis started in 2008.

36

The Built & Human Environment Review, Volume 3, Special Issue 1, 2010

As the detailed analysis of the data captured through the interviews has to wait until full transcripts are codified and the in-depth questionnaire is still expecting responses, the information generated so far is in the progress in addressing the overall aims and objectives of the research. The information gathered so far is indicative of how the apportionment of liabilities in delay claims is practiced in the local context, and existence of many problematic situations associated with such practices. This information is expected to be used towards developing a framework for improvement, including the targeted model for selecting the optimum delay analysis method for particular project circumstances. With reference to trustworthiness and validation of research findings, Miles and Huberman (1994) proposed a series of approaches in the form of questions for assessing goodness of research. Lincoln and Guba (1985) suggested four basic questions based on truth value, applicability, consistency and neutrality which must be addressed in the validity of any research inquiry. They suggested credibility, transferability, dependability and conformability as appropriate alternative criteria together with techniques such as triangulation, peer debriefing, discrepant evidence, thick description and audit trail. Besides these, there are more conventional criteria such as internal and external validity, reliability, objectivity and so on which are commonly used in quantitative approach. Thus, due to the mixed methods being used in data collection and analysis it was found more appropriate to use those approaches in the validation of the data analysis and the findings of this research. A part of the said in-depth questionnaire provides necessary data for the targeted model. The required information is captured through the practitioners perceptions, experience and application of delay analysis and such collected data will be analysed and used through the Analytic Hierarchy Process (AHP). In relation to the selected research strategy, the paper now proceeds to discuss on the relevance and use of AHP for developing the targeted model.

Analytic Hierarchy Process (AHP) and Its Significance in the Research

In delay claims resolution and apportioning liabilities, application of delay analysis methodology has an essential role in establishing the causation by technical proof opposed to proof by inference. There are such methodologies based on critical path method as well as other not so sophisticated methods of calculation. They are numerous and referred to by different names. The outcomes produced by some methods are more reliable and relatively objective, but those by others are overwhelmingly subjective. However, none of these methods is considered perfect. Ng et al. (2004) argue that the amount of delay attributable to the client and the contractor depends on the delay analysis technique employed. Bubshait and Cunningham (1998) argue that no one delay analysis technique suits every situation. SCL Protocol (2002) recognised the absence of such universally acceptable single method of analysis. AACE International Recommended Practice No.29R-03 (2007) points out that no forensic schedule analysis method is exact. Thus, the industry accepts that there is no single analysis methodology universally available and applicable to all situations of claims resolution. Selection of a methodology for forensic delay analysis relies on professional judgment and expert opinion of the practitioners. Therefore, it requires many subjective decisions and depends on number of criteria. These judgments invariably vary from one analyst to the other, as they are subjective, qualitative and impressionistic. The current research inquiry conducted so far has found that this imprecise situation on what technical approach to be used to quantify effects of delay causes problematic situations and disputes between the practitioners of both sides. It is noted that many disputes arise due to intuitive decision making in regard to analysis method. Intuitive decisions are not supported by data and documentation and may appear arbitrary. For intuitive decisions
37

The Built & Human Environment Review, Volume 3, Special Issue 1, 2010

it is difficult to get it accepted by others, particularly because decision maker is unable to justify it with persuasive logic (Saaty, 2006). The research study aims to bring about contractually sound, practical improvements to problematic situations in contemporary practices in delay claims resolution. Thus, an area of such improvements is identified in developing a model in order to aid the practitioners for the selection of a defendable and most appropriate delay analysis method under the specific circumstances of a project. The literature review of academic works found similar effort in the recent past. Bubshait and Cunningham (1998) conclude that the best analysis method is to be selected on the basis of time and resources available and the accessibility of project documentation. Braimah and Ndekugri (2007) identified eighteen factors in six groups that influence the selection of delay analysis methods, and ranked them according to their relative importance. Arditi and Pattanakitchamroon (2006) identify four criteria (data requirements, time of analysis, capability of methodology and time and effort required) for selection of delay analysis methods. SCL Protocol (2002) identifies seven factors that largely dictate the use of a method of analysis after completion of a project. However, relative to the other areas of delay claims that have been subject to research there is a dearth of empirical research study on the issue of selection of delay analysis method. During the preliminary inquiry and interviews stages, it is revealed that practitioners in the UAE experience substantial problems and dispute escalation during claims resolution due to individual, arbitrary approaches taken in the selection of analysis methodology. Thus, there is a significant need for further research in this area. It is in this context of decision making, the Analytic Hierarchy Process (AHP) is considered as a suitable and reliable tool for developing the targeted model.

History, Philosophy, and Mechanics of AHP

AHP belongs to the general family of Multiple Attribute Decision Making (MADM) methods. MADM refers to making preference decisions such as evaluation, prioritisation, and selection over the available alternatives that are characterised by multiple, usually conflicting, attributes (Hwang and Yoon, 1981). AHP was developed in 1970s by Thomas Saaty. Its philosophy is based on the intention to provide a comprehensive and rational framework for structuring a decision problem, for representing and quantifying its elements, for relating those elements to overall goals, and for evaluating alternative solutions. It is based on mathematics and psychology. Also, AHP is a normative model of decision making. Opposed to a descriptive model which allows describing the way a DM actually makes decision, a normative model enables a DM to defend his choice over competing alternatives. Presenting and clarifying the essential concepts of AHP for decision making process, Saaty (2008) submits that the measurement of concepts changes from one situation to another and they depend on a value system which varies from one person to another. Thus their importance cannot be measured once and for all and has to be determined in terms of individual values. The most appropriate way to measure such attributes is by comparing their relative importance with respect to a higher goal. From the comparisons, one derives a scale of priorities which are in relative values. Judgments come first, priorities second. In its performance, AHP refers to alternatives, attributes and goals. The term alternative is synonymous with option, policy, method etc. The term attributes may be referred to as objectives or factors or criteria. In any problematic situation attributes are multiple, and in the problem resolving process a DM generates relevant attributes for each problem setting. Keeney and Raffia (1976) suggested the use of a literature survey and /or a panel of experts to identify the attributes in the problem area. Yoon and Hwang (1995) stressed the necessity that attributes represent the desired mission and suggested one way to derive the attributes hierarchically from a super goal; goal hierarchy formulation starts with the listing of overall performance objectives serving a super goal. They noted such hierarchy to consist of at least three levels: focus or overall goal

38

The Built & Human Environment Review, Volume 3, Special Issue 1, 2010

at the top, multiple criteria that define alternatives in the middle, and competing alternatives at the bottom. The current research uses a similar hierarchy of attributes (i.e. criteria and sub-criteria) identified through literature review and interviews with practitioners, and a super goal of selection of optimum delay analysis method under given circumstances of a project. It is advised to limit the major and sub attributes in number seven based on Millers theory that seven plus or minus two represents the greatest amount of information that an observer can give about an object on the basis of an absolute judgment (Yoon and Hwang, 1995). This approach is followed in the research inquiry (through a questionnaire) seeking practitioners judgment over the importance over each attribute. While the number of attributes depends on the nature of the problem, each attribute may have different units of measurement. These units may be quantitative (e.g. number, money and so on) or qualitative (e.g. significance, necessity of presence and so on). Attribute weights are the quantitative way to express the importance of each attribute relative to the others, which is usually supplied and assessed by an ordinal or cardinal scale. The current research questionnaire-survey has supplied a cardinal scale based on a five point Likert-type-scale to assess such relative importance of attributes. As questionnaire-respondents judge and weigh such importance differently, the weights assigned by them for each attribute is required to be averaged prior to calculating attribute weighting. This weighted-average for an attribute can be expressed mathematically as: Wi x Xi /N (where Wi is the weight assigned to ith option, Xi is the number of respondents who selected the ith option; N is the total number of respondents.)
G

C1

C2

C3

A1

A2

Figure1. The simple three level-hierarchies below illustrate a typical AHP process

In this decision hierarchy structure (Figure1), at Level 1 the focus is the achieving Super-goal G. Level 2 comprises the Criteria C1, C2 and C3 which contribute to Super-goal. Each of these criteria is mutually independent and contributes to Super-goal differently. In this hierarchic arrangement of elements, the only restriction is that any element in one level must be capable of being related to some elements in the next higher level, which serves as a criterion for assessing the relative impact of elements in the level below (Saaty, 2006).

Thus, in the above structure, the criteria (C1, C2 and C3) are measured by comparing their relative importance with respect to Super-goal (G). Likewise, the relative importance of each alternative (A1 and A2) is measured with respect to each criterion (C1, C2 and C3) at higher level. Attribute weighting for AHP can be obtained through a simplified Eigenvector prioritization method presented by Saaty (1980). (Saaty also presented a method through Eigenvector raising the pair-wise matrix to powers that

39

The Built & Human Environment Review, Volume 3, Special Issue 1, 2010

are successively squared each time and then calculating and normalising the row sums. However, in order to keep this paper uncomplicated, only the simplified Geometric Mean method is described.) According to Saaty (2009) the fundamental scale of the AHP is a scale of absolute numbers used to answer the basic question in all pair wise comparisons: how many times more dominant is one element than the other with respect to a certain criterion or attribute?. Based on this principle, first, a pair-wise comparison is used to obtain the relative importance among the three Criteria. Comparison is based on the suggested numbers expressing the degree of preference or importance as assigned by the individual or a group of participants using a Likert scale. In the case of group-participants, their preference numbers are averaged to obtain weighted average for each Criterion. The importance ratios between three Criteria with respect to Level 1 Super-goal can be stored in a (nxn) matrix whose typical element ajk represents the weight ratio of wj/wk. The remaining elements of the matrix are filled by employing the reciprocal property of the matrix: ajk =1/akj and ajj=1, for all j and k. The number of pair-wise comparisons required can be expressed as n(n-1)/2. Thus, for three Criteria there will be just 3 comparisons. The pair-wise judgment amongst the three Criteria with respect to Level 1 can be concisely shown in the Figure 2 as follows, presuming judgments of preference are C1=3, C2=2 and C3=4:

C1 C1 C2 C3
w1/w1 w2/w1 w3/w1

C2
w1/w2 w2/w2 w3/w2

C3
w1/w3 w2/w3 w3/w3 = 1 .66 1.33 1.5 1 2 .75 .50 1

Secondly, geometric mean of each row of the matrix is computed and then the resulting numbers are normalized:

Geometric Mean C1 C2 C3 (1x1.5x0.75) 1/3 = (.66x1x0.50) 1/3 = (1.33x2x1) 1/3 = 1.0400 0.6910 1.3855

Weight (normalized) 0.3337 = 0.2217 0.4446

Sum =3.1165

1.0000

Through normalization, computational problems caused by differing measurement units of the criteria are eliminated and obtaining comparable scales can be achieved. This would allow both inter-attribute and intraattribute comparison as the normalized ratings have dimensionless units. The larger the rating becomes the more the preference given. Linear type normalization is enabled by dividing the ratings of a certain criterion or attribute by its maximum value. Thus, for jth attribute normalized value of xij is: rij = xij/xj, i=1.,m; j=1,.,n (Where xj is the maximum value of the jth attribute. So, 0 rij 1; and the attribute is more favourable as rij approaches 1). Result is similar when vector normalization is used instead of linear type.

40

The Built & Human Environment Review, Volume 3, Special Issue 1, 2010

Then, similar to the above process, pair-wise comparison of the two Alternatives A1 and A2 at Level 3 with respect to each of three Criteria at Level 2 will be carried out. Presumed comparisons are:

For Criterion C1

A1

A2

A1 A2

w1/w1 w2/w1

w1/w2 w2/w2 =

1 3

.33 1

Geometric Mean

Weight (normalized)

A1 A2

(1x.33) 1/2 = (3x1) 1/2 =

0.5744 1.7320 =

0.2490 0.7510

Sum =2.3064

1.0000

For Criterion C2

A1

A2

A1 A2

w1/w1 w2/w1

w1/w2 w2/w2 =

1 2

.50 1

Geometric Mean

Weight (normalized)

A1 A2

(1x.50) 1/2 = (2x1) 1/2 =

0.7071 1.4142 =

0.3333 0.6667

Sum =2.1213

1.0000

41

The Built & Human Environment Review, Volume 3, Special Issue 1, 2010

For Criterion C3

A1

A2

A1 A2

w1/w1 w2/w1

w1/w2 w2/w2 =

1 .50

2 1

Geometric Mean

Weight (normalized)

A1 A2

(1x2) 1/2 (.50x1)

1/2

= =

1.4142 0.7071 =

0.6667 0.3333

Sum =2.1213

1.0000

Thus, the relative contributions (i.e. weights) amongst the two Alternatives toward the three criteria are computed as: C1 C2 C3

A1 A2

0.2490 0.3333 0.6667 0.7510 0.6667 0.3333

Computation of contribution of each Alternative to the Super-goal (Level 1) by aggregating the resulting weights vertically, is the final step. Accordingly, the overall priority for each of Alternative A1 and A2 is obtained by summing the product of the criteria weight and the contribution of the alternative with respect to that Criterion.

Figure 2. Priorities for Each Hierarchical Level (Source: Yoon and Hwang, 1995)

42

The Built & Human Environment Review, Volume 3, Special Issue 1, 2010

The resulting computations for each of the Alternatives are:

A1 = 0.3337(0.2490)+0.2217(0.3333)+0.4446(0.6667)= 0.4534

A2 = 0.3337(0.7510)+0.2217(0.6667)+0.4446(0.3333)= 0.5466

Accordingly, the best Alternative for achieving the Super-goal is through A2 which has the higher value.

The value of Alternative can be mathematically expressed as: n V(Ai)=Vi = wj vj (xij ), i=1,..m J=1 (Where V(Ai) is the value function of Alternative Ai and wj and vj (. ) are weight and value functions of Alternative xj , respectively.)

AHP in This Research

The research study has established its philosophical basis on Critical Realism. Ontologically, Critical Realism accommodates the existence of multiple subjective meanings within participants personal and social worlds. The inquiry methods (questionnaire surveys and interviews) adopted in this research were designed to gather such data on multiple judgments, approaches and methods of practitioners who may construct meaning in different ways even for the same occurrence or facts. Thus, to handle such data and their analysis towards building the targeted model, compatible data analysis techniques (non-parametric) are required. AHP features the recognition of multiple constructed realities in differences within perceptions, judgments, attitudes and practices amongst practitioners dealing with delay claims. The judgments and the like are intangible and they have to be first measured before can be used as variables. As Saaty (2009) argued what is most significant is that intangible can only be measured through expert judgment and only relative to the goals of concern in a particular situation. The AHP is a tool that supports the relative measurements of such intangibles. It allows for differences in opinion to develop a best construct. On the other hand, the main character of AHP lies in its ability to take into account essentially emotional factors or attributes in a decision making process. Thus, AHP fits into the role in deriving measurements out of such subjective and qualitative data for decision making (selection of the optimum delay analysis method. It helps to resolve conflicts in judgments and bring together different perspectives of different practitioners to choose the best of a set of Alternatives. Thus, AHP enables a Decision Maker (DM) to combine quantitative and qualitative data. Despite its heavy reliance on mathematical manipulation, AHPs support mixed or multiple methods in data collection and analysis. To this extent AHP is utilised in this research from the Critical Realism perspective. Accordingly, AHPs role in using both quantitative and qualitative data collection and analysis is expected to result in the phenomenon of apportioning liabilities in delay claims resolution being seen from different perspectives illuminated with mixed methods. This places the AHPs performance in full harmony with the philosophical position of this research design which promotes the use of mixed or multiple methods.

43

The Built & Human Environment Review, Volume 3, Special Issue 1, 2010

Thus, AHP has been found to be suitable and reliable tool for achieving the targeted model which is an essential part of the overall aims and objectives of the research study.

Figure 3 below illustrates how AHP is placed and utilised in building a model for selection of the optimum delay analysis method that suits to project circumstances.

Pilot Questionnaire Literature Review Semi-structured Interviews

Identify Alternative analysis methods in practice

Identify Criteria and Subcriteria for selection of analysis methods

Design of in-depth Questionnaire to capture practitioner input

Data Collection

Analytical Hierarchy Process

Level 1 Goal

Decision Makers (DM) input at analysis process based on actual circumstances of rpoject

Rank - significance of Level 2 Criteria with respect to Level 1 Goal

Rank - significance of Level 3 Sub-Criteria with respect to Level 2 Criteria

Rank - significance of attributes of each Alternative with respect to Level 3 SubCriteria

Rank - degree of actual (physical) presence of Level 3 Sub-Criteria to enable implementation of each Alternatives

Establish alternative method of analysis having the highest (optimum) suitability rank

Optimum Analysis Method for the Project

Figure 3. AHP in building a Model for selection of the optimum delay analysis method

44

The Built & Human Environment Review, Volume 3, Special Issue 1, 2010

Conclusion and Further Research

The paper has discussed the concepts and selection of the Analytic Hierarchy Process (AHP) in the research, its ability to embrace both quantitative and qualitative data through assignment of numerical values to qualitative data, and the significance of its role as a decision making tool for building the targeted model aimed by the research. A simple three level-hierarchy is used to illustrate the basic AHP process. Prior to discussing the AHP, the paper discussed about Critical Realism as the research philosophy, and its ontological, epistemological and methodological positioning relative to realism (objectivist) and relativism (subjectivist). It also examined the issues in the research area from the Critical Realism perspective in consideration of the three defining questions of ontology, epistemology and methodology. As to data collection, Critical Realism approach led to use mixed-method studies where both quantitative and qualitative approaches are adopted for method of inquiry. This approach is expected to provide not only a basis for triangulation but also a source to look at the same things from different points of view of conceptualising the problem situation. It is observed that the performance of AHP is in full harmony with the philosophical position of this research design which promotes the use of mixed or multiple methods. The paper also described where the current stage of inquiry is. It concisely reviewed the findings of a pilot survey of which the findings have been used to conclude the need of carrying out this research and to formalise its aims and objectives. Following the pilot survey, a series of detailed interviews with the practitioners and the experts is in progress at the time of preparing this paper. Full transcripts of these interviews are being prepared in order to be codified under themes and sub-themes and then to be fed into the AHP model. Following this, an indepth questionnaire survey is also progressing at the moment to capture the practitioners perceptions, experiences, attitudes, judgments etc. embedded in delay claims resolution. The response period for this questionnaire is to be closed very shortly. Upon the completion of the data collection, the research will proceed to the next stage of analysing and interpreting the empirical data. Findings from the data analysis including the development of the AHP model will be reported in subsequent publications.

References
AACE (2007), International Recommended Practice No.29R-03, www.aacei.org (accessed on 28 October 2008). Archer,M., Bhaskar, R., Collier,A., Lawson,T., and Norrie, A. (Eds.). (1998), Critical Realism: Essential Readings, Routledge, London. Arditi, D. and Pattanakitchamroon, T. (2006), Selecting a delay analysis method in resolving construction claims, International Journal of Project Management, 24 (2006) pp145-155, www.elsevier.com/locate/infoandorg (accessed on 01 July 2010) Braimah, N. and Ndekugri, I. (2008),Factors influencing the selection of delay analysis methodologies, International Journal of Project Management, 26(2008), pp 789-799, www.sciencedirect.com (accessed on 10 July 2010) Bhaskar, R. (1978), A realist theory of science, Harvester Press, Hamel Hempstead . Bubshait A.A. and Cunningham, M.J (1998),Comparison of delay analysis methods. Journal of Construction Engneering Management, ASCE, 1998;24 (4), pp 315-322 Burrell, G. and Morgan, G. (1979), Sociological Paradigms and Organisational Analysis, Heinemann, London.

45

The Built & Human Environment Review, Volume 3, Special Issue 1, 2010

Creswell, J.W. (2007), Qualitative Inquiry and Research Design: Choosing among Five Approaches (2nd edition), Sage Publication, London. Crotty, M. (1998), The Foundation of Social Research, Sage Publication, London. Denzin, N.K. and Lincoln, Y.S. (1994), Handbook of Qualitative Research, Sage Publication, London. Easton,G. (2009),Critical Realism in Case Study Research, Industrial Marketing Management, 39(2010)118128, www.elsevier.com/locate/infoandorg (accessed on 01 July 2010) Groff, R. (2004), Post positivism and the Possibility of Knowledge Routeledge, London. Guba, E.G. and Lincoln, Y. S. (1994), Competing Paradigms in Qualitative Research, in: N.K. Denzin and Y.S. Lincoln (ed.), Handbook of Qualitative Research, Sage Publication, London. Guba, E.G. and Lincoln, Y. S. (1989), Fourth Generation Evaluation, Sage, California. Hwang, C.L. and Yoon, K. (1981), Multiple Attribute Decision Making: Methods and Application, SpringerVerlag, New York. Keeny, R.L. and raiffa, H (1976), Decisions With Multiple objectives, Wiley, New York. Lincoln, Y. S. and Guba, E.G. (1985), Naturalistic Inquiry, Sage Publication, London. Maxwell, J.A. (2005), Qualitative Research Design: An Interactive Approach (2nd edition), Sage Publication, London. Miles, M.B. and Huberman, A.M. (1994), Qualitative Data Analysis: An expanded Source Book (2nd edition), Sage Publication, London. Mingers, J (2004), Critical realism and information systems: brief responses to monod and Klein, Information and Organization, 14(2004) 145-153, www.elsevier.com/locate/infoandorg (accessed on 30 June 2010) Mingers, J. and Brocklesby, J. (1997),Multimethodology: Towards a Framework for Mixing Methodologies, Omega, Int.J.Mgmt Sci., Vol.25, No.5,pp 489-509, www.elsevier.com/locate/infoandorg (accessed on 30 June 2010) Ng. S.T, Skitmore, R.M., Deng, M.Z.M., and Nadeem, A. (2004),Improving delay analysis techniques for the establishment of delay liabilities, Construction Innovation 4(1): pp3-17, http://eprints.qut.edu.au/archive/00004116 (accessed on 18 July 2010) Patton, M. Q. (2002), Qualitative Research & Evaluation Methods (3rd edition), Sage Publication, London. Polkinghorne, D.E. (1989), Phenomenological research methods, in: R.S. Valle and S. Halling (ed.), Existential-phenomenological perspectives in psychology, Plenum Press, New York. Robson, C. (2002), Real world Research (2nd edition), Blackwell Publishing, Oxford. Saaty, T.L. (1980), The Analytical Hierarchial Process, Wiley, New York. Saaty, T.L. (2006), Fundamentals of Decision Making and Priority Theory with the Analystic Hierarchy Process (2nd edition), RWS Publications, Pittsburgh. Saaty, T.L. (2008), Relative Measurement And Its Generalization In Decision Making Why Pairwise Comparisons Are Central In Mathematics For The Measurement of Intangible Factors The Analytic Hierarchy/Network Process, RACSAM., Vol.102(2), 2008, pp 251-318, www.elsevier.com/locate/infoandorg (accessed on 11 July 2010) Saaty, T.L. (2009), Theory and Applications of the Analytic Network Process Decision making with benefits, opportunities, costs, and risks, RWS Publications, Pittsburgh.

46

The Built & Human Environment Review, Volume 3, Special Issue 1, 2010

SCL (2002), The Society of Construction Law - Delay and Disruption Protocol, www.scl.org.uk www.eotprotocol.com (accessed on 10 September 2007)

and

Tashakkori, A. and Teddlie, C. (1998),Mixed Methodology: Combining The Qualitative and Quantitative Approaches, Sage Publication, California. Yin, R.K. (2003), Case Study Research: Design and Methods (3rd edition), Sage Publication, London. Yoon, K.P. and Hwang C.L. (1995), Multiple Attribute Decision Making An Introduction, Sage Publication, California.

47

The Built & Human Environment Review, Volume 3, Special Issue 1, 2010

48