Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Abstract: In construction, delay claims resolution is at the center of apportioning liabilities between the claiming party and the defending
Downloaded from ascelibrary.org by Queens University Libraries on 12/04/16. Copyright ASCE. For personal use only; all rights reserved.
party. This process generally consists of two phases of causation: (1) establishing each party’s potential liability for the claimed occurrence;
and (2) determining the quantum of the effect flowing from that liability. However, a lack of consensus among the practitioners of disputing
parties as a result of the existence of various theories, concepts, and methodologies for apportioning liabilities has been acknowledged as an
obstructing factor in delay claims resolution. When determining the quantum of the effects flowing from a liability, one potential main
obstacle in apportioning liabilities is the existing dichotomous use of method of delay analysis (MDA), which can be responsible for
the inconsistent outcomes of delay analyses and result in distrust among the disputing parties. Aiming to address this problem, the research
project was set to develop a decision making model (DMM) to support practitioners in this matter. Following its validation, the developed
DMM enhances fairness and transparency in apportioning liabilities, hence overcoming possible skepticism about the impartiality of the
chosen MDA. Addressing the current need for an objective way to select the MDA, this DMM has made a contribution to the body of
knowledge of construction engineering, particularly in the claims management aspects, by providing a novel tool that can be considered
objective, reliable, and defendable for selecting the optimum MDA under the given circumstances of a construction project, and therefore
minimizing the potential of conflict and disputes in apportioning delays’ liabilities. DOI: 10.1061/(ASCE)ME.1943-5479.0000441. © 2016
American Society of Civil Engineers.
Author keywords: Delay analysis; Apportioning liabilities; Decision making.
Introduction Thus, delay claims resolution can be considered central to the ap-
portioning of liabilities between the claiming party and the defend-
Construction projects inherently carry risk that can be managed, ing party. This process consists of two phases of causation:
i.e., avoided, mitigated, transferred, or accepted, but cannot be (1) establishing each party’s potential liability for the claimed oc-
ignored (PMI 2013; Latham Report 1994). In many cases, projects currence; and (2) determining the extent of the effect flowing from
often carry the risk of delayed completion (Keane and Caletka that liability. The parties’ agreement for the approaches to be used
2008). This is a major source of claims, conflict, and disputes. in both phases is of vital necessity for the success of delay claims
Although claims may incur additional time and expense to deal resolution. However, in many projects these approaches are left to
with its documentation (Alkass et al. 1996), the occurrence of con- be decided until delay claims start emerging. This situation poten-
flict and disputes is of great concern in the construction industries tially poses complication for the decision makers of both the claim-
of many countries (Yates 1998). In Australia, for instance, the total
ing and defending parties. A clear lack of consensus and uniformity
days lost to construction industrial dispute per 1,000 workers can
of methodology among the decision makers and in many cases re-
be as high as 900 work days in a year, which is significantly higher
liant on decision makers’ individual perceptions and judgments
compared with the average of other industries in Australia (ACA
have been reported as the main problems (AACE 2011). Individual
1999). Construction delay has been suggested as one of the most
judgments are generally intuitive and subjective, and with intuitive
recurring and common problems (Mahamid et al. 2012; Rugaishi
decisions it is difficult to gain acceptance by others, particularly
and Bashir 2014; Gündüz et al. 2013), and modern construction
projects commonly suffer from it (Braimah 2013). Delay is ac- when the decision maker is unable to justify it with persuasive logic
knowledged as the most common, costly, complex, and risky prob- (Saaty 2006). It has been suggested that fair and equitable appor-
lem in construction projects (Fawzy and El-adaway 2012). Such tioning of the parties’ liabilities may be fettered by many factors in
delays in construction and the resulting claims require conscious contemporary practices that are critical in deciding the outcome
resolution through establishing liabilities and fair apportioning. (Braimah and Ndekugri 2007; Perera and Sutrisna 2010).
In determining the quantum of the effects flowing from a liabil-
1
Forensic Claims Consultant and Chartered Surveyor, School of Built ity, one potential obstructing factor in apportioning liabilities is the
Environment, Univ. of Salford, Salford M5 4WT, U.K. (corresponding lack of a universally acceptable method of delay analysis (MDA)
author). E-mail: nihal.a.perera@gmail.com among practitioners (AACE 2011; Keane and Caletka 2008). As
2
Associate Professor, Dept. of Construction Management, School of literature and the findings of a previous in-depth research inquiry
Built Environment, Curtin Univ., Western Australia 6102, Australia. posit (Perera 2012; AACE 2011; Keane and Caletka 2008), the
3
Senior Lecturer, Dept. of Civil and Environmental Engineering, Univ. existing dichotomous use of MDAs was found to be primarily
of Auckland, Auckland 1145, New Zealand.
responsible for inconsistent outcomes of delay analyses and sub-
Note. This manuscript was submitted on May 12, 2015; approved on
January 11, 2016; published online on April 6, 2016. Discussion period sequently distrust and skepticism among the disputing parties. In
open until September 6, 2016; separate discussions must be submitted delay claims resolution, application of MDA has an essential role
for individual papers. This paper is part of the Journal of Management in establishing the causation by technical proof, as opposed to proof
in Engineering, © ASCE, ISSN 0742-597X. by inference. Farrow (2001) pointed out that although decision
partiality of the particular method chosen. This study contributes correlating the planned program with actual progress. TIA-based
to the body of knowledge in the field of construction engineering analysis provides the consequences of an event on the assumption
management by providing a robust tool to assist practitioners in that future (remaining at the time of occurrence of the event)
deciding the most appropriate MDA, and therefore minimizing activities of the program will proceed as planned (as originally in-
the potential for conflict and disputes in apportioning delays’ liabil- tended). TIA is an evolution of the IAP method, and the difference
ities in the construction industry. between the two is the use of multiple baseline programs in TIA as
opposed to a single baseline in the IAP (AACE 2007; Keane and
Caletka 2008). Although it is very time consuming and expensive
Review of MDA in Apportioning Liabilities to implement (El-adaway et al. 2014), for many practitioners, TIA
is the most desirable approach to handle a delay claim, as long as
The general rules as to the recovery of damages (Schwartzkopf data and source documents are available in the required format and
and McNamara 2001) can be considered almost similar in the at the required time frame (SCL 2002; Baram 1994), despite its
U.K. and U.S. legal systems [e.g., Hadley v. Baxendale (1854) limitation of imprecise representation of the actual delay (SCL
9 Ex 341, 156 ER 145; Victoria Laundry (Windsor) Ltd., v. 2002, Section 4.8; Fruchtman 2000) or the fact that it does not scru-
Newman Industries Ltd., (1949) 2 KB 528; U.S. v. J. H. Copeland tinize the delay type before the analysis (thereby requiring further
& Sons [1928] 568 F.2d 1159; S. J. Groves & Sons Co., v. Warner analysis to apportion entitlement). Moreover, the concurrent delays
Co., (1928) 576 F.2d 524; Maxwell v. Schaefer (1955) 112 A.2d are not immediately addressed, which necessitates further analysis
69; R. J. Lampus Co., v. Neville Cement Products Corp. (1977) 378 that may become too cumbersome if there is an overwhelming
A.2d 288; Roanoke Hospital Association v. Doyle & Russell, Inc., amount of delaying events (Alkass et al. 1996).
(1975) 214 S.E.1d 155; Rochey Bros., Inc., v. Rhoades (1975) 527
F.2d 891; S. J. Groves & Sons Co., v. Warner Co., (1928) 576 F.2d
Collapse-as-Built (CAB)
524). According to these general rules, the complexities involved in
apportioning liabilities in construction delays are to be addressed The CAB method generally relies on a simulation of a what-if sce-
only through establishing causation (i.e., establishing the nexus nario, based on the critical path method (CPM). It models not the
between the cause of the delaying event, and the quantum of delay intentions of the contractor, but rather his or her actual sequences
effects flowing from that cause). Generally, for determining such and durations, by using a deductive approach that is exactly the
quantum of delay effects, scholars (e.g., Keane and Caletka 2008; opposite philosophy of an additive approach, which is relied on
Livengood 2007b; Farrow 2001; Baram 2000) have identified four by the IAP and TIA methods (Keane and Caletka 2008). The theory
primary MDAs: impacted-as-planned (IAP), time impact analysis of the CAB is that once the owner’s delays are subtracted from the
(TIA), collapse-as-built (CAB), and as-planned versus as-built actual project duration, it would reveal how long the contractor
(APvAB). would have actually taken to complete the project without the own-
er’s delay (Fruchtman 2000; Keane and Caletka 2008; Bubshait and
Impacted-as-Planned (IAP) Cunningham 1998). In its common form, CAB creates an as-built
schedule and identifies actual delays caused by one party, and then
IAP takes the as-planned schedule and adds new activities that removes them from that as-built schedule, to collapse the schedule.
generally represent the other party’s delays (contractor’s or employ- The argument for CAB is that without these delays, this is when
er’s, depending on who uses the method), to demonstrate why the the project would have been completed (Zack 2001). Associated
project was delayed beyond the completion date (Zack 2001). The limitations of CAB include the high level of subjectivity, difficul-
theory in IAP is that when all of the owner’s changes are added to ties in identifying as-built (contemporaneous) critical path, failure
the contractor’s original plan, the earliest date when the project to calculate delay based on the contractor’s contemporaneous in-
could have been completed because of the added work can be de- tentions (“at the time”) (Keane and Caletka 2008), and its inability
termined (Fruchtman 2000; Keane and Caletka 2008; Bubshait and to make allowance for mitigation of delays (Lovejoy 2004). It is
Cunningham 1998). Lucas (2002) and Bubshait and Cunningham suggested that CAB be used only when reliable as-built informa-
(1998) observed that although the IAP approach was widely used tion is available (El-adaway et al. 2014). Further discussions on
and accepted in the 1970s and early 1980s, it can now be rejected CAB can be found in the literature (e.g., Zack 2001; Fruchtman
by the courts primarily because of the subjectivity involved 2000; Al Saggaf 1998; Alkass et al. 1996; Schumacher 1995).
(e.g., Gulf Contracting, ASBCA No. 37,939, 94-2 BCA Para26.726
[1994]). The IAP has been generally criticized for not accounting
As-Planned versus As-Built (APvAB)
for changes to logic or durations of planned activities, producing
theoretical results based on hypothetical question, and the inability The APvAB method compares the actual and planned durations of
to identify true concurrent delay (Keane and Caletka 2008). It is the project and claims the difference as an excusable/compensable
based on a schedule that does not reflect actual progress of the delay. It does not require CPM logic and can be used by comparing
a graphic comparison of the as-planned schedule to the as-built particular chosen methodology. Table 1 has summarized the typical
schedule, simply observing any variances between the start/finish factors in selecting MDA
dates of the various activities. In other words, it resembles a total In delay claims resolution, one critical aspect that a decision
claim (Zack 2001), in which the claimant is required to discharge a maker may confront is the need to select an optimum MDA among
similar burden of proof as in the case of global claims (i.e., in which the MDAs used in contemporary practice. If such selection is based
the contractor does not seek to prove precisely his or her loss from on an intuitive and subjective approach, this would most likely in-
an event, but pursues a claim for the total loss suffered from all of vite further disputes over the selected MDA and its outcome.
the events). Accordingly, when its conventional version is used, the Although there is no generally accepted basis for what constitutes
contractor has to demonstrate that (1) the baseline program was a good decision (Harrison 1975), a decision can still be categorized
reasonable and realistic to succeed; (2) there were no contractor’s to the extent of its appropriateness, timeliness, effectiveness, and
critical delays contributing to the duration overrun; and (3) the time efficiency (Friday-Stroud and Sutterfield 2007). Thus, an appropri-
overrun (difference between the duration of as-planned and that of ate decision-making model is expected to satisfy these character-
as-built schedules) is entirely attributable to the delays of the em- istics, to support appropriate decisions on the selection of the
ployer or for which he or she assumes responsibility. To calculate optimal MDA.
the standalone effects of each of these delay events is impossible or The use of simple and yet sufficiently accurate techniques
impractical, as they are too complicated and intertwined to separate. (Ginevicius and Podvezko 2004) would allow the decision makers
In any case, considering its limitation to deal with concurrency and to use the decision-making model with better understanding and
dynamism of critical path, the results produced by this MDA may confidence than the one built on complex algorithms that are too
be contentious because an apportionment of delay liability may re- complicated to comprehend. In this paper, the authors intended to
sult in a flawed outcome if the effect of concurrent delays and choose one of the most widely used, simple methods like the ana-
changes in the critical path is overlooked (Alkass and Golanaraghi lytic hierarchic process (AHP) (Saaty 1980, 2006, 2009) or the sim-
2012). ple additive weighting (SAW) method (Hwang and Lin 1987;
Hwang and Yoon 1981; Yoon and Hwang 1995). Following an
Selection of MDA in-depth literature review on the matter, SAW was determined to
be the most appropriate method for this research. Table 2 lists some
Reflecting the multiple principles of MDA in practice as discussed research findings on the applications of SAW and AHP.
in the previous section, various scholars have recommended differ- In deciding between AHP and SAW, the main focus was the
ent factors when selecting MDA. Bubshait and Cunningham (1998) need for adopting a straightforward and simple method to capture
concluded that selection should be based on the time/resources the practitioners’ decision-making process. This input process
available and the accessibility of project documentation. Braimah involved numerous criteria and attributes to be judged; hence, it
and Ndekugri (2007) identified 18 factors that influence the selec- requires a simple but robust mechanism. The SAW method em-
tion of MDA. Arditi and Pattanakitchamroon (2006) identified four ploys additive weighted preferences, whereas AHP uses a ratio
factors toward the selection of MDA. SCL (2002) identifies seven scale to bring out pair-wise comparisons to collect data in devel-
factors that largely dictate the use of an MDA. In discussing the oping the system. Although both AHP and SAW are capable of
choice of a forensic schedule analysis methodology, AACE (2011) providing a robust algorithm, AHP typically requires a large num-
emphasized that, because individuals generally work for one party ber of pair-wise comparisons or judgments, whereas the SAW ap-
in a dispute, skepticism often exists regarding the impartiality of the proach requires much fewer direct and individual rankings. Thus,
the rationale for SAW’s main attraction to the practitioners over the Role of Attributes
other multiple attribute decision making (MADM) methods is its
In the problem-resolving process it is generally required to generate
simplicity, which makes it extremely popular in practice (Hobbs
the relevant attributes for each problem setting. This can be
et al. 1992; Zanakis et al. 1995, 1998).
achieved by reviewing literature and/or conducting interviews
(Keeney and Raiffa 1976). Yoon and Hwang (1995) stressed the
SAW Method in Research Perspective necessity that attributes represent the desired mission, and sug-
gested a way to derive the attributes hierarchically from a super
The basic principle of SAW involves establishing a weighted sum goal (as goal hierarchy formulation starts with the listing of overall
of mathematical performance ratings of each alternative under all performance objectives serving a super goal). Such hierarchy is ex-
attributes (MacCrimmon 1968; Chen and Hwang 1992). It works in pected to consist of at least three levels: (1) focus or overall goal at
two primary steps. First, it scales the values of all attributes, making the top, (2) multiple criteria that define alternatives in the middle,
them comparable; then the total score for each alternative is com- and (3) competing alternatives at the bottom. It is suggested that the
puted by multiplying the comparable value of each attribute by the basic principle of SAW is to obtain a weighted sum of the perfor-
importance weight assigned to the attribute and summing up these mance ratings of each alternative under all attributes (MacCrimmon
results over all of the attributes (Hwang and Yoon 1981; Virvou 1968; Chen and Hwang 1992).
and Kabassi 2004; Chou et al. 2008). This calculation process The weighted average of each criterion and attribute can be
of the value of an alternative can be submitted in the following mathematically expressed as
mathematical expression: Pn
wx
WAi i¼0 i i ð2Þ
X
n n
VðAi Þ ¼ wj rij i ¼ 1; 2; 3; : : : ; n ð1Þ
j¼0 where wi = weight (rate) assigned to the jth option; xi = number
of respondents who selected the ith option; and n = number of
respondents).
where VðAi Þ = value function of the ith alternative; rij = compa-
rable rating or normalized value of xij (i.e., the rated value of the ith
alternative for jth attribute); and wj = importance weight of the jth Research Methodology
attribute.
The alternative with the highest score of value function is ranked To capture these multiple perceptions and practices in resolving
the best option. delay claims, the research inquiry adopted a mixed method ap-
SAW has the ability to recognize multiple constructed realities proach. Mixed methods research is an approach to inquiry that
in differences within perceptions, judgments, attitudes, and practi- combines or associates both qualitative and quantitative data and
ces among practitioners dealing with delay claims. The judgments analysis. The use of qualitative and quantitative approaches and
and the like phenomena are intangible, and they have to be mea- the mixing of both approaches concurrently or sequentially in a
sured first before they can be used as variables. SAW can be used particular way that the overall strength of a study is greater than
readily as a tool or a method of relative measurements of such in- either the qualitative or quantitative research (Creswell and Plano
tangibles. It allows for differences in opinion with an ability to de- Clark 2007). In this case, the qualitative data (gathered from semi-
velop a best compromise. Using the Likert scale, such judgments structured interviews) provided detailed understanding of a prob-
can be converted into numerical values, giving the advantage of lem that arises out of studying few individuals and exploring their
comparison of such subjective data and information on a uniform perspectives in depth. In contrast, quantitative data (collected from
basis (Bowling 1997; Burns and Grove 1997; Likert 1932). Thus, it the questionnaire survey) provided a more general understanding of
enables an objective quantification of emotional factors or attrib- a problem that stemmed from examining a relatively large number
utes in a decision-making process. Accordingly, SAW fit into of people and assessing the response to the proposed variables.
the role in deriving (quantitative) measurements out of such sub- Thus, as suggested by Reams and Twale (2008), the mixed method
jective and qualitative data for decision making (i.e., selection of approach was found to be particularly appropriate in uncovering
the optimum MDA). It helps to resolve conflicts in human percep- information and perspective, increasing corroboration of the data,
tions and judgments, while bringing together different perspectives and rendering less biased and drawing more accurate conclusions.
of different practitioners to choose the best of a set of alternatives. Data from the semistructured interviews were analyzed using a
Thus, SAW enabled the combining of quantitative and qualitative content analysis technique (Weber 1990), and the data collected
data through assignment of numerical values to qualitative (subjec- from the questionnaire survey directly formed the SAW decision-
tive) data, and firmly displayed its ability to use mixed methods. making model. For validation purposes, the case-study approach
were drawn from the Royal Institution of Chartered Surveyors equation (Waller 2010; Sutrisna 2004; Cheung et al. 2009) was
(RICS) and Chartered Institute of Arbitrators (CIArb) membership considered:
database in the United Arab Emirates (UAE, particularly Dubai)
members, complemented by personal industry contacts. The practi- Z2 × P × ð1 − PÞ
SS ¼ ð3Þ
tioners (who were actually involved in delay claims resolution) C2
came from contractors, consultants/project administrators, develop-
where SS = sample size; Z = Z value (Z value = 1.96 for 95% con-
ers, independent claims consultants, and construction law back-
fidence level); P = maximum variance occurred, 0.5 providing the
ground. Even though the data collection was conducted using
worst-case scenario; and C = margin of error.
practitioners in the UAE (Dubai), the practices in UAE follow
Therefore
(and therefore can be considered comparable to those in) practices
in other jurisdictions, primarily in the United Kingdom and the 1.962 × 0.5 × ð1 − 0.5Þ
United States. In addition, the nature of the UAE (Dubai) construc- 74 ¼ ; C ¼ 11.4%
C2
tion industry has attracted a significant number of international
practitioners working in the sector, thus bringing their international In the selected sampling approach, the number of eligible re-
perspectives, expertise, and ways of working. To a certain extent, spondents (74) has an estimated margin of error of 11.4 at 95%
therefore, the UAE industry embodied similar constructs that are confidence level.
perceived and practiced by the practitioners in other comparable Based on their actual involvement revealed in the responses to
settings. the subsequent questions, these 74 respondents were then stratified
into two groups (Table 3)—the contracting group and the consult-
ing group—as the main competing parties in delay claims.
Results and Discussion Ten follow-up semistructured interviews were then conducted,
100 min in length each, then audio-recorded and fully transcribed
Initially, 520 potential respondents across the industry in Dubai/ with the permission of the interviewees. The content analysis of
UAE were identified through the snowball sampling technique the interview transcripts resulted in six major themes and 27 sub-
and circulated through the web-linked survey-questionnaire. This themes. The interviews were conducted and purposively selected
selection was based on their career association with construction using primarily the snowballing technique until data saturation
claims. However, over 200 e-mails were returned as not deliverable, was achieved (Glaser 1978; Goulding 2005).
All of these interviewees can be considered experts in delay
claims resolution, representing both sides of the fence (i.e., contrac-
Table 3. Composition of Two Groups tor and consultant/owner groups). Most of them were claims and
delay analysts with civil engineering and quantity surveying back-
Cumulative
Group Frequency Percent percent
ground, with a single case of a practicing construction lawyer (refer
to Table 4).
Contracting group 33 44.6 44.6 After analyzing both the qualitative (captured primarily through
Consultants/owners group 41 55.4 100.0 the interviews) and quantitative (obtained primarily from the sur-
Total 74 100.0 —
vey) data, the convergent or divergent findings were summarized,
program IAP, or APvAB is high but not for CAB, which does not require a baseline program)
6. Availability of CPM as-built • This may be necessary for some MDAs, but not for others (e.g., it is essential for TIA but not for APvAB, which
program updates can be implemented without frequently updated progress schedules)
7. Existence of other records • This is generally preferred by all MDAs, but for some they are essential (e.g., unless an as-built program is
mutually agreed by parties, CAB requires that its as-built critical path based be built on the other records)
8. Establishing causation • In general, for establishing causation for events in isolation, TIA, CAB, and IAP provide more convincing results
than APvAB
9. Establishing concurrency • If properly implemented, all MDAs are capable of establishing concurrency /mitigation, but IAP’s outcome is
more theoretical and subjective than others
10. Establishing responsibility • In general, CAB can directly illustrate isolated effects attributed to a particular delay event, and its source of
responsibility requires no further exercise; however, the other MDAs require a further round of evaluation to do
this
11. Cost of the delay analysis • IAP and APvAB are relatively cheaper to implement; however, when issues are complicated, use of TIA or CAB
is necessary, which may result in higher cost (for want of better expertise)
12. Time required/allowed to carry • In general, TIA and CAB take longer than IAP and APvAB
out delay analysis
13. Admissibility by triers-of-fact • Admissibility by arbitrators or judges may be higher for the MDAs like TIA and CAB than the other MDAs
compared, discussed, interpreted, and finally merged to see how attributes in a ranking order. Tables 6 and 7 list the two groups
the quantitative results build or expand on the qualitative findings, of respondents to the survey, who attribute the significance/
and how the merged results answer the research questions and importance to each of the criteria at level 2 and the attributes
objectives. at level 3 of the proposed DMM, respectively. In both Tables 6
and 7, the test statistics (Kendall’s W and Critical W) show with
considerable confidence that the agreement or concordance among
Development of the Decision-Making Model
the rankings of the respondents of the two groups is higher than it
The formulation of the DMM was based on a thorough literature would be by chance, had the scores and their rankings been selected
review undertaken during the research project. This literature randomly or independently.
review identified the selection factors used in this DMM and pro- The use of the data in the development of the SAW model is
vided conceptualization for the model as summarized in Table 5. described in the next section.
Using the findings from the survey questionnaire, as tabulated in The hierarchy of the proposed DMM is composed as
Tables 6 and 7, the DMM was further developed and refined. follows:
1. Level 1: The super goal (i.e., selection of the optimum method
of delay analysis);
Elements of the Decision-Making Model 2. Level 2: The criteria to achieve the super goal (the 7 criteria that
The current decision-making model uses a hierarchy of criteria, appear in Table 6 are identified, as generally informed by the
attributes (or subcriteria), and alternatives; all of these elements literature review);
have been identified through the literature review and the findings 3. Level 3: The attributes or subcriteria of these criteria (the 23
of the research. The purpose of the proposed DMM (or its super attributes that appear in Table 7, identified as they emerged
goal) is the selection of optimum MDA under the given circum- from the criteria and largely informed by the literature review,
stances of a project. The selection of criteria is driven by the ability as discussed previously); and
and influence to achieve the super goal; likewise, the selection of 4. Level 4: The alternatives or methods of delay analysis; these
attributes is made as they emerge from each of the elements of are the most widely used MDAs (i.e., as-planned versus as-
criteria in the level above. Although the number of criteria and built, impacted-as-planned, collapsed-as-built, and time impact
attributes depends on the nature of the problem, each of them analysis).
may have different units of measurement. These units may be quan-
titative (e.g., number, money) or qualitative (e.g., importance,
significance, necessity of presence). The use of assigned weights
Application of SAW
are intended to indicate the importance of each criterion/attribute
relative to the others. Based on the hierarchy formulation of the SAW method, scores
The survey questionnaire provided a measuring scale based on (rates) for the alternatives were to be obtained by adding the
a five-point Likert scale, to assess the relative importance of the contribution from each strand of attribute/criterion. However, the
ratings from the questionnaire respondents were based on differ- where m = number of alternatives as n is the number of attributes;
ent measurement units using a five-point Likert scale. Because xij = rated value of the ith alternative for the jth attribute; and xj =
different measurement units of items could not be added, a common maximum rated value scored by any of the alternatives for the jth
numerical scaling system, such as normalization, was to be used attribute.
to permit additions among the attribute values. Accordingly, the Then, 0 ≤ rij ≤ 1, and xij is more favorable as rij approaches 1.
scores of both criteria and attributes were normalized, and such Thus, each alternative is rated against 23 attributes and 1 cri-
normalization was applied on the weighted averages of their aggre- terion directly (criterion titled “legal admission” has no attributes
gated scores computed from the ratings given by the respondents. or subcriteria).
This was according to the literature review, which suggested that The value function of an alternative in the SAW method can be
the basic principle of SAW is to obtain a weighted sum of the per- mathematically expressed as
formance ratings of each alternative under all attributes.
Having established the weighted average of each criterion X
n
and attribute, the algorithm for the operation of the DMM was de- VðA1 Þ ¼ wj rij i ¼ 1; : : : ; m ð5Þ
veloped (using Microsoft Excel) in five major steps according to j¼1
the SAW principles. These five steps are given in Table 8.
The normalized value of xij can be presented as where VðAi Þ = value function of the ith alternative; and
rij = comparable rating or normalized value (see Step 3) of
xij ; and wj = importance weight (see Step 2) of the jth
rij¼ðxij =jÞ i ¼ 1; : : : ; m; j ¼ 1; : : : ; n ð4Þ attribute.
Step 4 • Compute the value functions of each alternative to select the highest scoring alternative
Step 5 • Select the alternative with the maximum AVF; it is the optimal MDA that suits the specific circumstances of the project
Among them were five fellows and one member of Chartered results out of 32 comparisons was tested using the exact binomial
Institute of Arbitrators, and all of them possessed Masters of probability test (Lowry 2008).
Law degrees in construction law and arbitration specialty. All nine Using the Microsoft Excel 2007 binomial distribution
members of the panel used the DMM in their own case studies for analysis formula [BINOMDIST (21,32,0.5,0.5)], Pð≤ 21Þ ¼
checking its performance under varying scenarios, confirming a fair 0.9749487701.
amount of interaction with the DMM. Thus, the probability of having 21 or more similar results from
In the outcome of the evaluation, 89% of the experts confirmed 32 relative comparisons was Pð≥ 21Þ ¼ 1 − ðP < 21Þ ¼ 1 − ½Pð0Þþ
that the content of the model was simple, comprehensive, and clear Pð1Þþ · · · þPð20Þ ¼ 1 − 0.9749487701 ¼ 0.0250212299.
for the user (with 67% stating it was “very simple, comprehensive This result indicates that the probability of having this level of
and clear”). Although during the model development stage several consistency in performance occurring by chance was significantly
iterations of the DMM were conducted to ensure a smooth running low. It confirms that the anomalies between the predicted and actual
of the final product, the panel of experts was also expected to use it outcomes of the DMM’s performance in each possible combination
and evaluate the accuracy of the DMM. According to the results, were significantly minimal. These replications also plausibly dem-
most (89%) of the panel decided that “the process is stable and runs onstrate that the model’s consistency in rational outcomes (causal
smoothly with no abnormal behavior.” The model’s representative- relationship between input and output) was highly satisfactory, and
ness of the breadth of domain knowledge was also an important thus having a sufficient internal validity.
aspect of the evaluation by the experts. The results of this evalu- Accordingly, the DMM demonstrated a reasonably high level
ation showed 86% of the experts stating it was “highly adequate” of objectivity, reliability, appropriateness, timeliness, effectiveness,
and 14% stating it was “reasonably adequate.” As to the scales of and efficiency.
scoring used in the DMM, the panel confirmed satisfaction, stating
they were “very suitable” (67%) or “generally suitable” (33%).
Consistency of the model’s delivery of the optimum MDA (output) Potential Limitations
under varying scenarios/circumstances (input) was found to be The main potential limitation of the DMM stems from the input of
stable and corresponding to the intended application, when tested the respondents to the in-depth survey. The process of weighted
through several iterations during the development stage of the averages and their normalized weights converted the raw data
model. The panel of experts also requested to test this aspect for collected from the respondents’ ratings (of relative “significance”/
determining whether the DMM yielded a satisfactory range of ac- “importance” among the criteria and attributes) into a dimension-
curacy consistent with intended application (Sargent 2003). The less and comparable common numerical scaling system. However,
results of this test showed an overall confirmation from the panel the original ratings were entirely determined by the individual
of experts that the changing scenarios (“input” based on varying levels of perception and experience of the respondents, and the
case-study scenarios) applied to the model generated a consistent competence of their judgments. Therefore, there is a high degree
and accurate output (the optimum MDA in relation to the “input”). of reliance toward the database of experts, albeit a significant level
Of them, 67% thought “output is very consistent and accurate for of concordance between the rankings of both groups of survey re-
the intended purposes of the model.” spondents, as indicated by the test’s statistical tests.
For testing internal validity, several iterations of the model Another potential limitation that may affect the final outcome
(runs) were carried out with varying potential scenarios in construc- are the parameters/factors and weighted ratings allocated for the
tion projects. The results proved a high consistency rate, showing decision maker’s input. The currently fixed parameters and ratings
that all “actual” and “predicted” results in 21 times out of a total related to decision makers’ input are proven to have reflected
32 comparisons were similar. The probability of having 21 similar reasonably meaningful behaviors of the MDAs with reference to
Fig. 1. Application of the decision-making model in the case study at levels 1, 2, and 3
This paper discusses the development and anatomy of the DMM method in resolving construction claims.” Int. J. Project Manage.,
24(2), 145–155.
that was specifically designed to enable the practitioners to choose
Baram, G. E. (1994). “Delay analysis—Issues not for granted.” Transac-
the optimum MDA on a more objective and tenable basis for delay
tions of AACE, Morgantown, WV.
analysis. It is an outcome of a recently completed doctoral research Baram, G. E. (2000). “The window methods of analyzing delay claims.”
project. AACEI Trans.
The development of the DMM was based on the simple additive Bernstein, S., and Bernstein, R. (1999). Schaum’s outline of theory and
weighting method as its decision-making mechanism/algorithm. problems of elements of statistics II inferential statistics, McGraw-Hill,
The SAW method uses a simplified arithmetical process. Its New York.
hierarchical components consist of a set of criteria and attributes Bowling, A. (1997). Research methods in health, Open University Press,
(or subcriteria) of which the relative significance/importance was Buckingham, U.K.
determined by a group of credible practitioners who participated Braimah, N. (2013). “Understanding construction delay analysis and the
in the research project. Data collection was facilitated by semistruc- role of preconstruction programming.” J. Manage. Eng., 10.1061/
(ASCE)ME.1943-5479.0000216, 04014023.
tured interviews (qualitative data) and an in-depth questionnaire sur-
Braimah, N., and Ndekugri, I. (2007). “Factors influencing the selection
vey (quantitative data). The mixed-method approach was adopted to of delay analysis methodologies.” Int. J. Project Manage., 26(8),
capture these practitioners’ perceived judgments, which was then 789–799.
captured in a database of experts to develop the DMM. The resulting Bubshait, A. A., and Cunningham, M. J. (1998). “Comparison of delay
DMM is capable of taking the input of a decision-making event and analysis methods.” J. Constr. Eng. Manage., 10.1061/(ASCE)0733
provided a recommendation in selecting the most optimum MDA -9364(1998)124:4(315), 315–322.
based on the specific characteristics of the particular project. Burns, N., and Grove, S. K. (1997). The practice of nursing research con-
The DMM has been validated through a series of case studies duct, critique, and utilization, W. B. Saunders and Co., Philadelphia.
and also evaluated by an expert panel (the details of which will be Burton, L. J., and Mazerolle, S. M. (2011). “Survey instrument validity.
published in subsequent publications). The findings of the reliabil- Part I: Principles of survey instrument development and validation in
athletic training education research.” Athletic Training Educ. J., 6(1),
ity and validation significantly confirmed that the developed DMM
27–35.
performs consistently to the purposes with an appropriate level of Chang, Y. H., and Yeh, C. H. (2001). “Evaluating airline competitiveness
accuracy. As evaluated by an expert panel, with a higher level of using multiattribute decision making.” Omega, 29(5), 405–415.
consistency rate, the DMM demonstrated an ability to provide an Chen, S. J., and Hwang, C. L. (1992). Fuzzy multiple attribute decision
objective and tenable basis for selecting an optimum MDA, com- making: Methods and applications, Springer, New York.
pared with any other intuitive and one-sided approach for such se- Cheung, S. O., Chow, P. T., and Yiu, T. W. (2009). “Contingent use of
lection. Therefore, it can be concluded that the research has negotiatiors’ tactics in construction dispute negotiation.” J. Constr. Eng.
developed a robust DMM to improve objectivity in the process Manage., 10.1061/(ASCE)0733-9364(2009)135:6(466), 466–476.
of selecting an optimum MDA to aid and yield a fair and equitable Chou, S. Y., Chang, Y. H., and Shen, C. Y. (2008). “A fuzzy simple additive
outcome in apportioning liabilities. Thus, this research paper con- weighting system under group decision-making for facility location se-
lection with objective/subjective attributes.” Eur. J. Oper. Res., 189(1),
tributes to the body of knowledge of construction engineering, and
132–145.
particularly the claims management community, by providing a re- Creswell, J. W., and Plano Clark, V. L. (2007). Designing and conducting
liable, accurate, and objective tool that can be used in minimizing mixed method research, Sage Publications, Thousand Oaks, CA.
the potential for conflict and disputes arising from apportioning li- El-Adaway, E. H., Fawzy, S. A., Bingham, R., Clark, P., and Tidwell, T.
abilities in construction delays. (2014). “Different delay analysis techniques applied to the American
Nevertheless, because the main mathematical basis for the func- Institute of Architects A201-2007 standard form of contract.” J. Legal
tion of the DMM is largely determined by psychological constructs Affairs Dispute Resolut. Eng. Constr., 10.1061/(ASCE)LA.1943-4170
of the individuals (respondents), further research is recommended .0000145, 02514001-1–02514001-9.
for ensuring continuous refinement and expansion of the DMM Farrow, T. (2001). Delay analysis—Methodology and mythology, 〈http://
in its database of experts by incorporating them into the DMM’s www.scl.org.uk〉 (Sep. 2, 2008).
Fawzy, S. W., and El-Adaway, E. H. (2012). “Contract administration
database.
guidelines for effectively and efficiently applying different delay analy-
sis techniques under world bank- funded projects.” J. Legal Affairs Dis-
pute Resolut. Eng. Constr., 10.1061/(ASCE)LA.1943-4170.0000104,
References 35–44.
Friday-Stroud, S. S., and Sutterfield, J. S. (2007). “A conceptual framework
AACE International. (2007). Recommended practice No. 29R-03, 〈http:// for integrating six-sigma and strategic management methodologies
www.aacei.org〉 (Oct. 28, 2008). to quantify decision making.” TQM Mag., 19(6), 561–571.
AACE International. (2011). Recommended practice No. 29R-03, 〈http:// Fruchtman, E. Z. (2000). “Delay analysis—Eliminating the smoke and
www.aacei.org〉 (Jun. 18, 2012). mirrors.” AACE Int. Trans., Cdr.06.1–Cdr.06.4.
Houghton Mifflin, Boston. theory with the analystic hierarchy process, 2nd Ed., RWS Publica-
Hobbs, B. J., Chankong, V., Hamadeh, W., and Stakhiv, E. (1992). “Does tions, Pittsburgh.
choice of multicriteria method matter? An experiment in water resource Saaty, T. L. (2009). Theory and applications of the analytic network
planning.” Water Resour. Res., 28(7), 1767–1779. process—Decision making with benefits, opportunities, costs, and risks,
Hwang, C. L., and Lin, M. J. (1987). Group decision making under multi- RWS Publications, Pittsburgh.
ple criteria: Methods and applications, Springer, New York. Sargent, R. G. (2003). “Verification and validation of simulation models.”
Hwang, C. L., and Yoon, K. (1981). “Multiple attribute decision making– Proc., 2003 Winter Simulation Conf.
Method and applications.” A state-of-the-art survey, Springer, Schumacher, L. (1995). “Quantifying and apportioning delay on construc-
New York. tion projects.” Cost Eng., 37(2), 11–13.
Keane, P. J., and Caletka, A. F. (2008). Delay analysis in construction
Schwartzkopf, W., and McNamara, J. J. (2001). Calculating construction
contracts, Wiley-Blackwell, Oxford, U.K.
damages, 2nd Ed., Aspen Publishers, New York.
Keeny, R. L., and Raiffa, H. (1976). Decisions with multiple objectives,
SCL (Society of Construction Law). (2002). Delay and disruption protocol,
Wiley, New York.
〈http://www.scl.org.uk or http://www.eotProtocol.com〉 (Sep. 10, 2007).
Latham, M. (1994). Latham Rep.—Constructing the Team, Her Majesty’s
Sutrisna, M. (2004). “Developing a knowledge based system for the
Stationery Office, London.
valuation of variations on civil engineering works.” Ph.D. thesis, Re-
Likert, R. (1932). “A technique for the measurement of attitudes.” Arch.
search Institute in Advanced Technologies, Univ. of Wolverhampton,
Psychol., 140, 1–55.
Wolverhampton, U.K.
Livengood, J. C. (2007a). The new AACEI recommended practice on foren-
Taylor, B., Sinha, G., and Ghoshal, T. (2008). Research methodology: A
sic schedule analysis—Part five of a five-part series: Choosing the right
guide for researchers in management and social sciences, Mudrak,
methodology, 〈http://www.aacei.org〉 (Apr. 7, 2008).
New Delhi, India.
Livengood, J. C. (2007b). The new AACEI recommended practice on foren-
sic schedule analysis—Part one of a five-part series: The methodology Virvou, M., and Kabassi, K. (2004). “Evaluating an intelligent graphical
debate, 〈http://www.aacei.org〉 (May 28, 2007). user interface by comparison with human experts.” Knowledge-Based
Lovejoy, V. A. (2004). “Claims schedule development and analysis: Syst., 17(1), 31–37.
Collapsed as-built scheduling for beginners.” Cost Eng., 46(1), 27–30. Waller, D. L. (2010). Statistics for business, Taylor and Francis, Hoboken,
Lowry, R. (2008). Concepts and applications of inferential statistics, NJ.
〈http://faculty.vassar.edu/lowry/webtext.html〉 (May 21, 2011). Weber, R. P. (1990). Basic content analysis, Sage, Newbury Park, CA.
Lucas, E. D. (2002). “Schedule analyzer pro—An aid in the analysis of Yates, D. J. (1998). Conflict and disputes in the development process: A
delay time impact analysis.” Cost Eng., 44(8), 30–36. transaction cost economics perspective, 〈http://www.prres.net/
MacCrimmon, K. R. (1968). “Decision making among multiple attribute proceedings/proceedings1998/Papers/Yates3Ai.PDF〉 (Nov. 26, 2014).
alternatives: A survey and consolidated approach.” RM-4823-ARPA, Yin, R. K. (1994). Case study research: Design and methods, Sage,
RAND Corp., Santa Monica, CA. Newbury Park, CA.
Mahamid, I., Bruland, A., and Dmaidi, N. (2012). “Causes of delay in road Yoon, K. P., and Hwang, C. L. (1995). Multiple attribute decision making—
construction projects.” J. Manage. Eng., 10.1061/(ASCE)ME.1943-5479 An introduction, Sage Publications, Thousand Oaks, CA.
.0000096, 300–310. Zack, J. G., Jr. (2001). “But-for schedules-analysis and defense.” Cost Eng.,
Modarres, M., and Sadi-Nezhad, S. (2005). “Fuzzy simple additive weight- 43(8), 13–17.
ing method by preference ratio.” Intell. Autom. Soft Comput., 11(4), Zanakis, S., Mandakovic, T., Gupta, S., Sahay, S., and Hong, S. (1995). “A
235–244. review of program evaluation and fund allocation methods within
Netemeyer, R. G., Bearden, W. O., and Sharma, S. (2003). Scaling proce- the service and government sectors.” Soc. Econ. Plann. Sci., 29(1),
dures: Issues and applications, Sage Publications, London. 59–79.
Perera, N., and Sutrisna, M. (2010). Research methodological position for Zanakis, S. H., Solomon, A., Wishart, N., and Dublish, S. (1998). “Multi-
a doctoral study on apportioning liability in delay claims, 〈http://www attribute decision making: A simulation comparison of select methods.”
.irbdirekt.de/daten/iconda/CIB18326.pdf〉 (Jul. 30, 2011). Eur. J. Oper. Res., 107(3), 507–529.