Case Study
Project Delivery Method Performance for Public School
Construction: Design-Bid-Build versus CM at Risk
Downloaded from ascelibrary.org by Charlie Roberts on 11/17/16. Copyright ASCE. For personal use only; all rights reserved.
Noel Carpenter, Ph.D. 1; and Dennis C. Bausman, Ph.D. 2
Abstract: The construction of public schools is a complex endeavor due to the involvement of multiple and diverse parties, funding and
budgetary concerns, and statutory limitations imposed by local, state, and federal agencies, all of which serve to increase project risk. Project
delivery methods using distinctive procedures to manage the design and construction process have been developed to reduce risk and improve
project performance. To increase the probability of successful project outcomes, those responsible for public school construction require
conclusive delivery method performance data. Completed in 2014, this two-year study used actual construction documents from 137
southeastern public schools to analyze and determine project delivery method cost, time, quality, and claims performance. The analysis
indicated that performance of the Design-Bid-Build (DBB) method was significantly superior across all cost metrics, whereas the Construction Manager at Risk (CM at Risk) method produced higher levels of product and service quality. Essentially, public school administrators
were paying a significant premium to obtain perceived improvements in both service and product quality. This research empowers decisionmakers and benefits the public by providing evidence of the most efficient and effective means for the construction of new public schools.
DOI: 10.1061/(ASCE)CO.1943-7862.0001155. © 2016 American Society of Civil Engineers.
Author keywords: Project delivery method; Design-Bid-Build (DBB); Construction Manager at Risk (CM at Risk); Project success
factors; Project performance factors; Contracting.
Introduction
Although all construction projects are subject to a wide variety of
complex issues and risks (Saporita 2006; Zaghloul and Hartman
2003; Akintoye and MacLeod 1997; Gordon 1994), the construction of public schools is particularly complex due to the diversity of
the parties involved, schedule and funding intricacies, and statutory
requirements (Vincent and McKoy 2008). Over the past 20 years,
the public education system has spent more than $174 billion on
new schools, and the median per-square-foot construction cost of
these facilities has doubled (Abramson 2013). During this same
period, budget shortfalls experienced at federal, state, and municipal levels have placed pressure on capital expenditures for public
schools, while growing populations and changing demographics
have served to increase demands on aging and outdated facilities
(Abramson 2012; Oliff et al. 2012; McNichol et al. 2011; U.S.
Census Bureau 2011).
Comprehensive management systems, known in the construction industry as project delivery methods, have been developed
to reduce project risk by assigning contractual responsibilities to
those involved in the design and construction process (Kenig
2011). The most widely used project delivery method is the
Design-Bid-Build (DBB) method (D’Agostino and Bridgers 2010).
However, proponents of alternative methods, such as Construction
Manager at Risk (CM at Risk) and Design-Build (DB), believe that
these methods offer the promise of better performance when used
1
Clemson Univ., 312F Lee Hall, Clemson, SC 29634 (corresponding
author). E-mail: [email protected]
2
Clemson Univ., 312F Lee Hall, Clemson, SC 29634. E-mail:
[email protected]
Note. This manuscript was submitted on October 21, 2015; approved on
January 12, 2016; published online on April 19, 2016. Discussion period
open until September 19, 2016; separate discussions must be submitted
for individual papers. This paper is part of the Journal of Construction
Engineering and Management, © ASCE, ISSN 0733-9364.
© ASCE
on certain types of projects (Cox et al. 2011; Konchar and Sanvido
1998). District managers and other public school administrators,
acting in the capacity of facility owners as guided by state and district procurement policies, are responsible for selecting the delivery
methods they believe are best suited to provide positive results for
their new school projects. This project delivery method selection
can often influence project success or project failure (Demkin
and AIA 2008).
Ghavamifar and Touran (2008) suggested that unnecessary
added costs, lack of experience with the processes, loss of owner
control, and the fear of favoritism associated with the awarding
of construction contracts may be influencing public decisions regarding the adoption and use of alternative methods of project
delivery. Additionally, anecdotal evidence suggests that state and
district agencies and their public school administrators and managers may be continuing to avoid alternative delivery methods
because of their lack of knowledge and experience or because
of traditional operating procedures currently in effect (Carolinas
AGC 2009).
Public school decision-makers require current, relevant, and significant delivery method performance data along with a thorough
working knowledge of the factors affecting school construction
to make informed, critical public school construction decisions
(Vincent and McKoy 2008). However, a review of the literature
revealed that only a limited amount of empirical research has been
conducted on project delivery method performance for the construction of public school projects. The empirical performance data
obtained through such research is required to assist decisionmakers in their efforts to select the most appropriate delivery methods for the construction of their schools. Therefore, the purpose of
this research was to provide current, statistically significant, empirical evidence to define the performance attributes of the traditional
DBB project delivery method as compared with the alternative
delivery methods of CM at Risk and DB when used for the construction of public school projects.
05016009-1
J. Constr. Eng. Manage., 2016, 142(10): 05016009
J. Constr. Eng. Manage.
Definitions
Although varying definitions exist within the construction industry,
the following descriptions obtained from the Associated General
Contractors 2011 publication, Project Delivery Systems for Construction by Michael E. Kenig, were used to define the project
delivery methods discussed within this research. The operational
definitions adopted for this study include the following:
Downloaded from ascelibrary.org by Charlie Roberts on 11/17/16. Copyright ASCE. For personal use only; all rights reserved.
Project Delivery Method
The comprehensive process of assigning contractual responsibilities for designing and constructing a project, which should include
the definitions of project scope, contractual responsibilities, interrelationships of the parties, and the processes for managing time,
cost, safety, and quality.
Design-Bid-Build (DBB)
The defining characteristics of this project delivery method are as
follows:
1. The design and construction are separate contracts: ownerdesigner and owner-contractor; and
2. The total construction cost is a factor in the final selection of the
constructor.
Construction Manager at Risk (CM at Risk)
The defining characteristics of this project delivery method are as
follows:
1. The design and construction are separate contracts: owner to
architect, owner to CM at Risk; and
2. The total construction cost is not a factor in the final selection
of the constructor.
Design-Build (DB)
The defining characteristic of this project delivery method is
that the design and construction responsibilities are contractually
combined into a single contract with the owner.
Public Schools
Public schools refer only to publicly funded school(s), and grades
kindergarten through twelfth grade (K-12).
District Manager
District manager refers to the person responsible for the construction of public schools at the district level (the facility owner).
of the multiple entities began to complicate the communication and
decision-making required during design and construction. The reduced cooperation and collaboration experienced among the parties
often causes the development of adversarial relationships and increases project risk, which may lead to reduced quality, schedule
overruns, change orders, claims, and litigation (Saporita 2006).
Contracting terminology and methodologies have been created
to define the project scope and assign responsibilities for completing the work to avoid, reduce, transfer, or maintain risk. Development of the DBB project delivery method occurred toward the end
of the nineteenth century following a number of fraud and abuse
cases associated with the transcontinental railroad and other large
U.S. infrastructure projects (U.S. DOT 2006; Heady 2013). Known
as the traditional method throughout the United States, DBB was
developed primarily to reduce the risk of corruption and cost overruns. It can be used to produce quality results on both public and
private projects when used under the proper conditions. Advantages associated with the DBB method include an easily understood
and well-documented process, the perception of fairness, owner
control of the process, reduced chances for corruption, reliable
schedule predictability, and initial cost certainty (Rojas and Kell
2008; Kenig 2011; U.S. DOT 2006). Disadvantages of this approach include limited opportunities for collaboration, development of adversarial relationships (which may lead to increased
litigation), and an increased probability of contractor failure as a
result of the competitive nature of the bidding process. Additionally, because of the linear structure of the design and construction
process, DBB is known to possess a restricted and expensive
process for incorporating constructive changes (Konchar 1997;
O’Connor 2009; U.S. DOT 2006; Cox et al. 2011).
Alternative delivery methods such as CM at Risk and DB seek
to close the gaps created by fragmentation by using collaborative
and communicative strategies in an effort to reduce both the probability of risk occurrence and the exposure to risk issues (Konchar
1997; Akintoye and MacLeod 1997; O’Connor 2009). Establishment of a team environment with collaborative exchanges of information and experience among the owner, architect, and contractor
throughout the project life-cycle has the ability to improve project
performance in terms of cost, schedule, quality, and productivity
(Konchar 1997; Konchar and Sanvido 1998; O’Connor 2009;
National Research Council 2009; Kenig 2011). Additionally, the
procurement of CM at Risk and DB projects is typically accomplished
using a more collaborative qualifications-based approach in lieu of
competitive bidding, further improving the project team relationships
(Kenig 2011). An added benefit of these methods is the opportunity
to begin construction of the project before the completion of final
design documents, which may improve schedule performance.
Project Success Factors
Literature Review
Project Delivery Methods
Prior to the Renaissance, most large-scale construction projects were
completed using a master builder approach, in which a single person
or entity would perform the duties of the architect, engineer, and
contractor (Addis 2007; Fitchen 1986; Ghavamifar and Touran
2008). Fragmentation of the master builder into the separate disciplines of architecture, engineering, and construction occurred during
the Renaissance and continued into the twentieth century, allowing
for the specialization of construction services required to support
technological advances and increasing complexity of building
projects. As a consequence of this separation, the disparate interests
© ASCE
Project delivery method selections are based on the personal experiences and organizational purchasing philosophies of those in
decision-making capacities (Sanvido and Konchar 1999). Research
indicates that the factors used to define project success are determined by the experiences and perceptions of the owner, which have
a direct influence on project delivery method selection (Chan and
Chan 2004). The factors used to define construction project success
have been the focus of many research efforts, but the first mention
of critical success factors (CSFs) comes from information systems’
research conducted at the Massachusetts Institute of Technology
(Rockart 1979). The definition was refined and presented by Bullen
and Rockart (1981) as “the limited number of areas in which satisfactory results will ensure successful competitive performance for
the individual, department, or organization. CSF’s are the few key
05016009-2
J. Constr. Eng. Manage., 2016, 142(10): 05016009
J. Constr. Eng. Manage.
Downloaded from ascelibrary.org by Charlie Roberts on 11/17/16. Copyright ASCE. For personal use only; all rights reserved.
areas where ‘things must go right’ for the business to flourish and
for the manager’s goals to be attained (Ibid).”
Although many indicators exist to gauge a wide array of performance factors, this research uses the CSFs most often used for
measuring construction project success: cost, schedule, and quality
(Atkinson 1999). Cost and schedule performance are seen as objective measures, whereas work quality (product quality) and client
satisfaction (service quality) are viewed as more subjective or intangible (Chan and Chan 2004).
The premise behind the development of the differing project
delivery methods is the result of the belief that each of them offers
different methods for managing and controlling the CSFs in an effort to improve project performance. For example, CM at Risk provides
an opportunity to reduce construction duration (an objective measure
of the schedule) by changing the linearity of the project life-cycle.
Additionally, client satisfaction (a subjective measure of quality)
may be improved by the perceived increase in cooperation and communication instilled by CM at Risks’ more collaborative properties.
Project Delivery Method Research
A limited amount of research has been performed to directly compare the performance of project delivery methods when used for the
construction of public schools. A report issued by the California
Council of the American Institute of Architects (AIA) cited the
“lack of quality data and information on school construction cost,
schedule, and scope, (and that) little research on these (construction) processes exists” (Vincent and McKoy 2008).
Foundational research conducted to compare the performance of
project delivery methods was conducted by Bennett et al. (1996),
Konchar (1997), and Konchar and Sanvido (1998); these studies
continue to be the most widely cited and accepted. Results of the
Bennett et al. (1996) research indicated that alternative methods,
such as DB and CM at Risk, performed better than DBB in terms
of cost and time. Univariate analysis conducted by Konchar (1997)
using median data indicated that the performance of DB and CM at
Risk projects exceeded the performance of DBB in terms of cost,
schedule, and almost all issues of quality. Konchar’s multivariate
analysis indicated that DB outperformed both CM at Risk and
DBB in terms of unit cost and construction speed (Ibid). In the
Konchar and Sanvido (1998) research, DB was shown to outperform both CM at Risk and DBB in terms of cost and schedule,
whereas quality performance results were mixed.
Findings of the Konchar (1997) and Konchar and Sanvido
(1998) works were challenged by Williams (2003) based on the
statistical significance of the results and wide variations of the comparative project types included in their study, which led to gross
disparities of overall project costs, square foot sizes, and complexity. The Williams (2003) research was directed at the comparative
performance of 215 Oregon public schools constructed with the
CM/GC (CM at Risk) and DBB methods using statistical analysis
and data envelope analysis (DEA). Results of the study indicated
that statistically significant differences did not exist in terms of cost
or schedule, although observable evidence revealed that schools
constructed using CM at Risk had a higher per-square-foot cost
than DBB schools. Additionally, the study revealed the CM at Risk
method may not provide the reduction in risk characteristics touted
by its supporters, and that no “interaction effect” (advantage based
on management strategy or tactics) was found for either method.
However, the study did indicate that CM at Risk may provide superior performance for projects that require accelerated (fast-track)
project schedules.
Additional research contradicting the CM at Risk performance
superiority findings provided by Konchar and Sanvido (1998) was
© ASCE
conducted by Rojas and Kell (2008). Focused on cost control performance during the construction of public schools in Oregon and
Washington, the study revealed only marginal, yet insignificant,
differences (1.55% mean, 0.48% median) between CM at Risk
and DBB in terms of change order costs, 75% of CM at Risk projects exceeding the Guaranteed Maximum Price (GMP), and CM at
Risk being significantly less effective at controlling costs during buyout (12.25% higher bid cost growth and 15.15% higher project cost
growth). Essentially, the results indicate that owners using CM at Risk
for construction of public schools experienced significantly greater
cost growth above the prebid estimates, without receiving lower
change order costs, nor the touted benefits of reduced risk and
lower cost volatility expected with a “Guaranteed” Maximum Price.
Methodology
This research was conducted during a 2-year period that began in
2012 and concluded in 2014. The study was conducted in the
southeastern states of Florida, Georgia, North Carolina, and South
Carolina, and included all public schools (K-12) that were completed between January 2006 and December 2012. The work
included a two-stage data collection effort, consisting of a historical
document review and assemblage followed by a survey of the district managers (owners) responsible for the construction of the
public school projects. The focus of the research was to determine
the actual design and construction costs of public schools, and to
define the comparative performance differences among the DBB,
CM at Risk, and DB delivery methods. The metrics that are developed to assess and compare project performance in terms of cost,
time, and quality are presented in the following sections.
Cost Metrics
Costs were limited to the design and construction costs associated
with the actual public school buildings and supporting sitework. To
the extent possible, costs for offsite roadwork, owner-furnished
equipment, and land were not included in the calculations. RS
Means (2013) factors were used to normalize all costs to 2012 dollars. The operational and maintenance costs of these facilities and
other life-cycle issues related to building performance were not
taken into consideration in this study. Therefore, a determination of
owner-perceived values obtained through the use of high-performance
materials or energy-efficient equipment was beyond the scope of
the research. The following cost variables were collected from the
project documents and then used to formulate the cost metrics
shown in Table 1:
• Original Contract Cost—the amount listed in the construction
contract and entered as Original Contract Sum on the Contractor’s Application for Payment;
• Preconstruction Cost—the amount listed in the construction
contract payable to the contractor for preconstruction services,
if applicable;
• Other Cost—the amounts listed on separate contracts including
sitework or other scopes and phases of work that were not
included in the Original Contract Sum, but were required to
complete the entire project scope;
• Final Contract Cost—the amount entered as the Contract Sum
to Date or other adjusted total amount listed on the Final
Contractor’s Application for Payment;
• Original Design Cost*—the amount listed in the design contract. In cases in which a percentage fee was listed, the amount
was calculated based on the architect contract terms; and
05016009-3
J. Constr. Eng. Manage., 2016, 142(10): 05016009
J. Constr. Eng. Manage.
Table 1. Cost Metrics
Metric
Formula
Original construction cost ($)
Original project cost ($)
Final construction cost ($)
Final project cost ($)
Construction cost growth (%)
Project cost growth (%)
Unit cost ($=m2 )
Student cost ($=student)
Original contract cost ($) + preconstruction cost ($) + other cost ($)
Original construction cost ($) + original design cost ($)
Original construction cost ($) + change orders, fees, adjustments costs ($)
Final construction cost ($) + final design cost ($)
½ðFinal construction costð$Þ − original contract costð$ÞÞ=original contract costð$Þ × 100
½ðFinal project costð$Þ − original project costð$ÞÞ=original contract costð$Þ × 100
Final project cost/facility gross square meters
Final project cost/facility student capacity
Downloaded from ascelibrary.org by Charlie Roberts on 11/17/16. Copyright ASCE. For personal use only; all rights reserved.
Table 2. Time Metrics
Metric
Formula
Planned construction (days)
Actual construction (days)
Planned project (days)
Actual project (days)
Construction growth (%)
Project growth (%)
Construction intensity (m2 =day)
Construction intensity ($=day)
Number of days between construction start date and original completion date
Number of days between construction start date and substantial completion date
Number of days between design start date and original completion date
Number of days between design start date and substantial completion date
½ðActual constructionðdaysÞ − planned constructionðdaysÞ=planned constructionðdaysÞ × 100
½ðActual projectðdaysÞ − planned projectðdaysÞÞ=planned projectðdaysÞ × 100
Facility gross m2 =actual projectðdaysÞ
Final project costð$Þ=actual projectðdaysÞ
Table 3. Quality Metrics
Metric
Formula
Product quality
Service quality
Service quality (readiness)
Product
Product
Product
Product
quality (readiness)
quality (readiness)
and service quality
and service quality
Owner satisfaction ratings of work quality for building exterior, building interior, environmental
systems, and the overall project
Owner satisfaction ratings of the design and construction team’s ability to manage and control
project costs, schedule, and product quality using methods of communication, collaboration, and cooperation
Number of days between the date of substantial completion and date of final completion (time to
complete all punch list and other miscellaneous items of work)
Number of warranty and callback incidents
Cost severity (to owner) of warranty and callback issues
Number of construction claim incidents
Cost severity (to owner) of construction claims
• Final Design Cost*—the total amount listed on the architect’s
Final Invoice.
*Separate Design Cost analyses were conducted in which reimbursable costs were both included and excluded in the Design Cost
calculations, which resulted in no significant differences between
the two. All known reimbursable costs have been included within
the research analysis and reporting.
Time (Schedule Duration) Metrics
The following time variables were collected from the project
documents, which were then used to formulate the time metrics
presented in Table 2:
• Design Start—date listed in the design contract or, in the case
that the date was not provided, the date the design contract was
executed;
• Construction Start—date listed in the construction contract or as
listed in the Notice to Proceed;
• Original Completion—substantial completion date listed in the
construction contract or in the Notice to Proceed, or computed
by adding the number of days listed in the construction contract
to Construction Start;
• Revised Completion—date that the Original Completion is
shifted to by adding (or subtracting) the number of days allowed
(or deducted) because of a change orders;
© ASCE
• Substantial Completion—date listed on the Certificate of
Substantial Completion; and
• Final Completion—date that the architect executed the
Contractor’s Final Application for payment.
Quality Metrics
The research challenge was in formulating metrics that properly
assessed the levels of product and service quality performance
and in crafting survey questions that would properly illicit participant perspectives regarding these subjective issues. Care was taken
to establish multiple metrics and targeted survey questions to
capture the true measures of product and service quality performance, while reducing variations that result from the subjectivity
of the respondents. The complete list of Quality Metrics is presented in Table 3.
Ratings of owner satisfaction with regard to public school product quality performance were obtained using survey data with
metrics directed at the workmanship of the building interior and
exterior, environmental systems, and the overall project. Service
quality satisfaction ratings were obtained using metrics focused
on the abilities of the design and construction teams to manage
and control project costs, schedule, and product quality using methods of communication, collaboration, and cooperation.
05016009-4
J. Constr. Eng. Manage., 2016, 142(10): 05016009
J. Constr. Eng. Manage.
Downloaded from ascelibrary.org by Charlie Roberts on 11/17/16. Copyright ASCE. For personal use only; all rights reserved.
Three empirical measures of building readiness were developed
as additional measures of quality performance. First, schedule data
were used to establish the number of days required by contractors
to complete all punch list and other miscellaneous items of work.
For this service quality metric, readiness was defined as the measure of days between the date of Substantial Completion and the
date of Final Completion (date the architect executed the Contractor’s Final Application for Payment). The second and third empirical measures of readiness were developed using records of warranty
and callback issues as indicators of product quality. The number of
warranty and callback incidents and their severity (in terms of cost)
were used to define these comparative measures of readiness.
The final two empirical measures of quality were developed
through analysis of owner-provided construction claims data. Often
caused by issues related to communication, project documentation,
contract terminology, relationships, productivity, and workmanship, construction claims can lead to increased costs and schedule
overruns (Saporita 2006; Akintoye and MacLeod 1997; Gordon
1994; Kangari 1995). Measures of the number of construction
claim incidents and their severity (in terms of cost) were used
as indicators of overall (product and service) quality performance.
Data Collection and Distributions
To reduce the variability among projects included within the study,
only new construction, full facility, public school projects were
targeted for data collection. The initial search for historical project
data was targeted at the departments of education within each
state included in the study area. As shown in Fig. 1, state records
contained a population of 829 public school projects that were
completed during the study period. Based on the population, calculations were completed to determine a minimum sample size of
90 projects and the proportional numbers of projects required from
each state to achieve valid results through statistical analysis.
Because of the limited and varied types of public school construction data collected and maintained at the state level, the historical
project data collection effort was focused directly at the more than
247 local school districts in which the public schools were constructed. All school districts were solicited for participation in this
study, and a concerted effort was focused on school districts with
active expansion programs, which correlated with those districts
having larger student populations.
Use of a survey approach for historical data collection purposes
proved to be ineffective in obtaining precise cost and schedule
information during early research efforts. Crosschecks of districtprovided data revealed that some respondents appeared to be
approximating the project cost and schedule information without
providing accurate figures. Therefore, it was determined that direct
reviews of actual project documents would be required to verify
and record accurate design and construction costs and schedule durations. Copies of the necessary documents were collected using
postal mail, email, and a number of on-site meetings during which
the researcher was able to review and copy the pertinent project
documents directly from district files. Complete sets of documents
that were assembled and reviewed for each project included the
owner contractor agreement, owner architect agreement, notice
to proceed, certificate of substantial completion, final construction
application for payment, final architect invoice/billing, and final
construction change order. This method of data collection improved
the accuracy of the data and the precision of the results achieved
during the data analysis stage. This is the first and only known
project-delivery-method performance research of this magnitude
that was conducted using actual construction project documents
for data collection and verification purposes.
The initial review of public school construction data within the
study area revealed that the DB method had only been used for new
construction of full-facility public school projects in the states of
Georgia and Florida during the study period. Furthermore, only
2 of the 286 projects completed in Georgia and only 15 of the
230 qualifying projects completed in Florida were constructed
using DB. Therefore, due to the limited use of the DB method
and the consequent limited amount of project data representing this
method, analysis of the performance of the DB method for public
school construction was not included within this research.
Once the historical data collection work was completed for a
particular project, an Internet-based survey was distributed directly
to the district manager (owner) responsible for the construction of
public schools within the individual districts. The survey was used
to obtain owner perceptions and ratings of the product quality of
the new facility and the quality of service provided by the construction and design teams during the design and construction process.
Because surveys were only distributed to managers in districts
where complete sets of actual project documents had been obtained
and personal discussions with respondents regarding the forthcoming survey had been conducted, a 93.3% survey response ratio was
obtained. Complete project delivery method performance data sets,
including information obtained through historical document reviews and surveys, were obtained from a sample of 137 proportionally distributed projects as shown in Fig. 1, which exceeded the
calculated minimum requirements previously discussed.
When distributed by project delivery method, the sample data
of completed projects reflect similar ratios to those of the U.S.
commercial sector. As presented in Fig. 2, the largest proportion
of projects (65%) were completed using the DBB method, with
a lesser portion (35%) being completed with CM at Risk.
However, when distributed by both project delivery method
and state as shown in Fig. 3, the data are clearly differentiated.
The project delivery methods used in North Carolina and South
Carolina resemble those of the population and the U.S. commercial
sector. Districts in Georgia predominantly use DBB, whereas those
in Florida overwhelmingly use the CM at Risk method.
Normalization and Testing Procedures
Before conducting the analysis, all costs were normalized to 2012
U.S. dollars using historical cost indexes and methods provided by
Population: N = 829 Projects
South Carolina,
109, 13%
North Carolina,
195, 24%
Sample: n = 137 Projects
Florida,
239, 29%
South Carolina,
24, 18%
Georgia,
286, 34%
North Carolina,
43, 31%
Florida, 30,
22%
Georgia,
40, 29%
Fig. 1. Population and sample distributions
© ASCE
05016009-5
J. Constr. Eng. Manage., 2016, 142(10): 05016009
J. Constr. Eng. Manage.
Sample: n = 137 Projects
US Commercial Sector (FMI/CMAA 2010)
Design-Bid-Build,
86, 63%
Design-Bid-Build,
55 , 55%
Other,
21 , 21%
CM at Risk,
24 , 24%
CM at Risk,
51, 37%
Downloaded from ascelibrary.org by Charlie Roberts on 11/17/16. Copyright ASCE. For personal use only; all rights reserved.
Fig. 2. Sample distribution by project delivery method
North Carolina Sample: nNC = 43
CM at Risk,
16, 37%
South Carolina Sample: nSC = 28
Design-Bid-Build,
27, 63%
Florida Sample: nFL = 51
CM at Risk,
26, 87%
Design-Bid-Build,
19, 79%
CM at Risk,
5, 21%
Georgia Sample: nGA = 43
Design-Bid-Build,
4, 13%
Design-Bid-Build,
36, 90%
CM at Risk,
4, 10%
Fig. 3. Sample distribution by project delivery method and state
RS Means (2013). Hypothesis tests and statistical evaluations
were completed using two group t-tests of the mean performance
values of cost, schedule, size, capacity variables, and chi-square
(x2 ) distributions for the quality-performance metrics. All testing
procedures were carried out based on 95% confidence intervals using the statistical analysis system (SAS/STAT version 9.3) software.
More than 600 individual tests were completed to compare the
mean performance values of the CM at Risk and DBB public
school projects.
An analysis of the responses to individual survey questions
was completed using chi-square (x2 ) distributions. Although the
lower two survey-response categories, “Very Dissatisfied” and
“Dissatisfied,” were selected more often by DBB managers than
by CM at Risk managers, very few responses were provided
within these two categories for any individual survey question
for either method, thus rendering analysis of these categories unviable for some questions. In most cases, the responses received in
the lowest two categories were combined into a single “Dissatisfied” category to improve the viability for statistical analysis. In
cases in which an adequate number of combined responses still
did not exist, the lower two categories were eliminated, leaving
only the highest two categories, “Satisfied” and “Very Satisfied,”
available for comparable analysis. Therefore, the quality performance Tables 7–10 present the combined total percentages of
owner responses provided in the “Satisfied” and “Very Satisfied”
categories.
© ASCE
Table 4. Size and Capacity Comparison of Design-Bid-Build and CM at
Risk
Metric
DesignBid-Build
CM at
Risk
13,999.56 13,386.32
Gross project
size (m2 )
Student capacity 1,073.66 1,041.78
Area per
12.93
13.14
student (m2 )
Difference Percentage P-Value
613.24
4.38
0.5766
31.88
−0.21
2.97
−1.62
0.6643
0.6570
Results
Comparative Size and Capacity Characteristics
Initial testing was completed by comparing the size and capacity
characteristics of the DBB and CM at Risk school projects, resulting
in no statistically significant differences in mean school project area
size, student capacity, or area per student ratios. As listed in Table 4,
the mean project size of the DBB projects were shown to be
613.2 m2 larger with a student capacity 31.8 students greater than
the CM at Risk projects. The project area per student of the DBB
and CM at Risk projects differed by a mere 0.2 m2 . Additional testing by project type (elementary, middle school, and high school)
05016009-6
J. Constr. Eng. Manage., 2016, 142(10): 05016009
J. Constr. Eng. Manage.
Table 5. Cost Comparison of Design-Bid-Build and CM at Risk
Metric
Downloaded from ascelibrary.org by Charlie Roberts on 11/17/16. Copyright ASCE. For personal use only; all rights reserved.
Original construction ($)
Final construct cost ($)
Original project cost ($)
Final project cost ($)
Unit cost ($=m2 )
Student cost ($/student)
Construction cost growth (%)
Project cost growth (%)
Design-Bid-Build
CM at Risk
Difference
Percentage
P-Value
$20,960,467
$21,280,286
$22,023,229
$22,427,554
$148.80
$20,915.60
1.25%
1.45%
$26,001,207
$26,101,221
$27,141,092
$27,436,298
$191.60
$27,057.00
0.32%
1.04%
($5,040,740)
($4,820,935)
($5,117,863)
($5,008,744)
($43)
($6,141)
0.93%
0.41%
−24.05
−22.65
−23.24
−22.33
−28.76
−29.36
74.40
28.28
0.0148
0.0230
0.0180
0.0250
<0.0001
<0.0001
0.2249
0.6004
and by state (Georgia, North Carolina, and South Carolina) revealed
no statistically significant project size, student capacity, or area
per student differences among the public schools being compared.
The Florida CM at Risk sample proved to be significantly larger in
both size and student capacity because of the inclusion of four K-8
schools in the middle school category. The differences were not
significant when the four oversized schools were removed.
Cost Performance Findings
Proponents of CM at Risk believe the collaborative properties and
other attributes of this project delivery method enable the construction of projects with better cost performance than those produced
using the DBB method. However, the analysis performed in terms
of cost performance for construction of public schools revealed that
projects constructed using the DBB method significantly outperformed those constructed with the CM at Risk method across all
cost metrics. As presented in Table 5, comparisons were completed
for the original and final construction costs and the original and
final project costs. The computations for project costs include design
fees. Comparisons of unit and student costs were computed using
school project size and student capacity means and final project cost.
Similar results were obtained when testing was conducted by
individual state and school type with significant differences found
in both North Carolina and Florida and for elementary and middle
schools. Mean Design Costs of CM at Risk projects were shown to
be 7.3% higher than fees for DBB projects, although the difference
was not statistically significant. However, Design Costs for North
Carolina CM at Risk projects were 51.5% higher, a significant difference with a P-Value of 0.0240. Furthermore, testing by delivery
method, state, and type combined resulted in a significantly higher
Design Cost mean for North Carolina CM at Risk elementary
school projects with a difference of $405,062 (56.4%) higher than
DBB with a P-Value of 0.0010. Unit and student costs of the North
Carolina elementary schools constructed with CM at Risk were also
shown to cost $503.77=m2 and $5,446.80=student higher than
public schools constructed using the DBB method. The analysis
indicated that the cost growth on CM at Risk school projects was
not significantly different from the cost growth on DBB projects.
Cost performance is not necessarily an indication of project
value. Projects employing high-performance materials, systems,
and methods may or may not achieve life-cycle cost benefits. Costs
associated with these life-cycle issues were not measured or included within this study.
Schedule Performance Findings
Proponents of CM at Risk believe the collaborative properties and
other attributes of this method, which include its ability to enhance
the schedule through the use of fast-tracking, may enable the construction of projects with better schedule performance than those
produced using the DBB method. However, the analysis indicated
no significant schedule performance differences between the public
school projects constructed with the CM at Risk and DBB project
delivery methods as presented in Table 6.
The observed differences in schedule growth appear to show
that the CM at Risk delivery method provides a higher level of
schedule predictability. Examination of the data revealed that differences are generated primarily within the original schedules and
increase through use of the change order process. Project intensity
(m2 =day) and project intensity ($=day) were used as measures of
productivity to determine the amount of project work (by area and
cost) put in place per school project (design and construction)
schedule day. Primarily due to the previously confirmed similarity
of size and schedule characteristics, the analysis indicated no statistically significant differences between CM at Risk and DBB
school projects with regard to project intensity (m2 =day). Moreover, although significant differences were found when testing
by project intensity ($=day), the usefulness of this metric as a measure of productivity performance was ineffectual, as the difference
is predominantly attributable to the higher cost of the CM at Risk
school projects. Additional tests of construction intensity, which
use the construction schedules in lieu of the project schedules, were
conducted and resulted in similar findings.
Quality Performance Findings
The results of the survey data analysis revealed that projects completed with CM at Risk significantly outperformed DBB projects in
all areas of product quality. As presented in Table 7, significantly
larger percentages of “Satisfied” and “Very Satisfied” responses
were provided by CM at Risk district managers with P-Value
≤0.0105.
Table 6. Schedule Comparison of Design-Bid-Build and CM at Risk
Metric
Actual construction (days)
Actual project (days)
Construction schedule growth (%)
Project schedule growth (%)
Project intensity (m2 =day)
Project intensity ($=day)
© ASCE
Design-Bid-Build
CM at Risk
Difference
Percentage
P-Value
569.00
1,023.60
15.58%
8.11%
14.74
$22,924.90
564.70
1,008.30
12.41%
6.52%
14.18
$28,784.40
4.30
15.30
3.17%
1.59%
0.57
($5,859.50)
0.76
1.49
20.35
19.61
3.84
−25.56
0.8930
0.8250
0.4868
0.4760
0.6084
0.0033
05016009-7
J. Constr. Eng. Manage., 2016, 142(10): 05016009
J. Constr. Eng. Manage.
Table 7. Owner Satisfaction with Construction Product Quality
Metric
Table 11. Owner Reported (%) Disputes and Claims
DesignCM
Bid-Build at Risk Difference Percentage P-Value
Overall quality
Building exterior
Building interior
HVAC system
Plumbing system
Lighting system
Warranty issues
93.02
86.05
90.70
96.52
93.02
93.03
93.02
100.00
100.00
100.00
98.00
100.00
98.00
95.74
6.98
13.95
9.30
1.48
6.98
4.97
2.72
7.50
16.21
10.25
1.53
7.50
5.34
2.92
0.0058
0.0018
0.0028
0.0105
0.0083
0.0017
0.0073
Metric
DesignBid-Build
CM
at Risk
90.70
9.30
0.00
94.12
5.88
0.00
Zero
1–3
More than 3
Downloaded from ascelibrary.org by Charlie Roberts on 11/17/16. Copyright ASCE. For personal use only; all rights reserved.
DesignCM
Bid-Build at Risk Difference Percentage P-Value
Metric
Overall performance
Planning
Cost control
Schedule control
Quality control
Communication
Cooperation
90.48
89.29
91.67
83.34
90.48
92.86
91.67
96.00
98.00
98.00
94.00
100.00
98.00
96.00
5.52
8.71
6.33
10.66
9.52
5.14
4.33
6.10
9.75
6.91
12.79
10.52
5.54
4.72
0.0040
0.0029
0.0007
<0.0001
<0.0001
0.0047
0.0083
Table 9. Owner Satisfaction with Design Team Service Quality
DesignCM
Bid-Build at Risk Difference Percentage P-Value
Metric
Capture owner vision
Complete documents
Clearly defined
documents
Timely responses
Communication
Cooperation
95.35
88.37
88.37
100.00
100.00
96.00
4.65
11.63
7.63
4.88
13.16
8.63
0.0562
0.0004
0.0149
91.86
91.86
91.86
93.88
95.92
94.00
2.02
4.06
2.14
2.20
4.42
2.33
0.0084
0.0532
0.0036
Table 10. Owner Satisfaction with Project Team Service Quality
Metric
Overall service
quality
Communication
Cooperation
Collaboration
DesignCM
Bid-Build at Risk Difference Percentage P-Value
90.70
96.00
5.30
5.84
0.0025
91.86
90.70
88.37
98.00
96.00
98.00
6.14
5.30
9.63
6.68
5.84
10.90
0.0864
0.0038
0.0039
As given in Table 8, the results indicate that the performance
of the CM at Risk method for construction of public schools was
superior to that of DBB in all areas of construction team service
quality, including overall service, planning, cost control, schedule
control, quality control, communications, and cooperation with
P-Value ≤0.0083.
Analysis of district manager survey responses (Table 9) indicates that CM at Risk projects significantly outperformed DBB
projects in all areas of design team service quality with the exception of capture owner vision and communication for which the
P-Value were marginally above the 0.05 level at 0.0562 and
0.0532, respectively.
As presented in Table 10, CM at Risk district managers responded with significantly higher satisfaction of project team
service quality in the areas of overall quality, cooperation, and collaboration when compared with responses from DBB managers.
Although an observable difference was obtained with project team
© ASCE
3.42
−3.42
0.00
Percentage
P-Value
—
—
—
3.77
−36.77
0.00
Table 12. Owner Reported (%) Warranty and Callback Issues and
Associated Costs
Metric
Table 8. Owner Satisfaction with Construction Team Service Quality
Difference
DesignCM
Bid-Build at Risk Difference Percentage P-Value
Zero to two
3–5
6 or more
$0
$1 to $5,000
More than $5,000
21.43
40.00
38.57
0.54
0.21
0.24
14.63
36.59
48.78
0.76
0.20
0.05
−6.80
−3.41
10.21
21%
−2%
−19%
−31.73
−8.52
26.47
39.27
−8.96
−79.91
0.5104
0.0405
Table 13. Service Quality and Readiness (Days)
Metric
Readiness (days)
Design-Bid-Build
CM at Risk
Difference
Percentage
P-Value
347.40
312.90
−34.50
−9.93
0.4708
communication, the P-Value of 0.0864 was above the chi-square
significance region of 0.05 set for the analysis.
The analysis indicated there were not enough data to conclude a
significant difference in the mean number of disputes and claims
between CM at Risk and DBB school projects. CM at Risk district
construction managers reported three projects (5.9%) with one to
three disputes and claims, whereas DBB managers reported eight of
their projects (9.3%) within the same range. Because of the relatively small number of disputes and claims reported, there were also
not enough data to produce a viable analysis regarding the cost difference for this issue. The full range of responses is provided in
Table 11.
Analysis of warranty and callback issues and their associated
costs were used as empirical measures of product quality readiness.
The analysis indicated no significant difference in the mean number
of issues occurring in CM at Risk and DBB school projects,
although a significant difference was found in the costs of these
issues. Even though the observed data show a higher number of
occurrences for the CM at Risk projects in Table 12, the evidence
indicates that the performance of the CM at Risk project delivery
method was significantly better at controlling the cost impact of the
warranty and callback issues that did occur.
The final empirical measure of quality was computed using
schedule performance data. As given in Table 13, no statistically
significant differences were found between the mean service quality readiness values of the CM at Risk and DBB projects included
in the study. Both methods required more than 300 days to complete all punch list and other miscellaneous items of work after the
date of substantial completion.
Conclusion and Recommendations
The impetus behind this research was to address the ongoing construction industry question: Which project delivery method, CM at
05016009-8
J. Constr. Eng. Manage., 2016, 142(10): 05016009
J. Constr. Eng. Manage.
Downloaded from ascelibrary.org by Charlie Roberts on 11/17/16. Copyright ASCE. For personal use only; all rights reserved.
Risk or DBB, performs at a higher level in terms of cost, time, quality, and claims? In so doing, this study focused on providing current, statistically significant, empirical evidence to define the
comparative performance attributes of both methods when used
for the construction of public schools. This research has revealed
the following empirical evidence:
• The cost performance of the DBB project delivery method was
significantly superior to that of the CM at Risk method for the
construction of public schools when comparisons were made
across all cost metrics;
• The product quality performance of the CM at Risk project
delivery method was significantly superior to that of the DBB
method for the construction of public schools when comparisons were made across all product quality metrics;
• The service quality performance of the CM at Risk project
delivery method was significantly superior to that of the
DBB method for the construction of public schools when comparisons were made across all service quality metrics, with the
exceptions of design team capture of owner vision, communication and project team communication, for which observable
evidence was obtained indicating CM at Risk performance
superiority; and
• Evidence was not obtained to support the superiority of either of
the two project delivery methods being compared in terms of
cost growth, time (schedule duration), time variance (schedule
growth), claims, or warranty and callback performance.
some districts have implemented the use of prototypical designs,
standardized equipment, and prequalification procedures to reduce
unpredictable results (risk). These actions may have resolved many
of the issues for which the collaborative properties of the CM at
Risk method are purported to provide an advantage, thereby making the projects more suitable for the DBB method. Furthermore,
the evidence indicated that CM at Risk contractors were sometimes
not hired early enough in the process to provide preconstruction
services of substantive value, which would have limited the effectiveness of CM at Risk. Additionally, although significant schedule
performance differences were not found, observable evidence did
indicate that the collaborative properties of the CM at Risk method
may have improved original schedule predictability and increased
the number of days awarded to contractors in the change order
process. Finally, it is theorized that the lack of schedule differences
found between the two methods may have had more to do with the
nature of public school construction schedule requirements (that
facilities be open and operable to meet a rigorous public participation schedule) than to the project delivery method being used.
Additional possible explanations for the cost, schedule, quality,
and risk performance differences experienced within this research
may be related to the selection of an improper project delivery
method, given the particular set of project design and construction
circumstances. Evidence was obtained showing that incorrect
owner perceptions and unrealistic expectations of project delivery
method abilities influenced the delivery method decision-making
process, which served to increase performance differences.
Implications
The literature has exposed many important issues within the construction industry including high levels of risk, project complexity,
and a lack of trust among the participants for which the CM at
Risk project delivery method is theoretically equipped to improve,
alleviate, or control.
District construction manager responses to survey questions focused specifically on collaboration and cooperation among the
individual and collective construction, design, and project teams
have provided evidence that the collaborative properties of CM
at Risk have enabled this method to perform at significantly higher
product and service quality levels than the DBB method. This evidence supports the foundational research and theoretical construct
that collaborative environments have the ability to positively influence the product and service quality of construction projects. Conversely, the cost performance results indicate that the collaborative
properties of the CM at Risk method were not capable of controlling or reducing costs, time, or risk (claims) on these projects. In
fact, costs for the CM at Risk projects were significantly higher.
Furthermore, virtually all survey responses to product and service
related questions were in the “Satisfactory” and “Very Satisfactory”
categories for both CM at Risk and DBB projects, indicating that
almost all schools within the study met the required standards and
expectations regardless of the delivery method being used. Essentially, public school administrators were paying a significant premium for superior, although perhaps superfluous, levels of service
and product quality.
The resulting performance differences from the foundational research and theoretical construct may be attributable in part to the
fact that the current research was focused strictly on public sector
projects that are similar in both size and type, as opposed to the
wide varieties and sizes of both public and private projects included
in the foundational research. Further explanation may be related to
project complexity. Although public school projects appear to be a
good fit for CM at Risk because of the number of involved parties
and their complex nature, evidence was obtained showing that
© ASCE
Recommendations
This research has shown that the collaborative properties of the CM
at Risk method provide significant beneficial product and service
quality performance improvements when used for constructing
public school projects. However, the evidence indicates that these
enhancements come with significant increases in the cost of construction. This study has also shown that the DBB method has
the ability to provide public school projects with satisfactory quality levels, and with the added benefit of significantly lower costs.
Furthermore, this research has cited previous studies confirming
the inherent risks within the construction industry and specifically associated with public school projects that serve to complicate the construction process. Therefore, those in decision-making
capacities will be required to make value-based decisions when
selecting the most appropriate delivery methods for the construction of public schools within their districts, to balance budgetary
concerns with their required product and service quality levels.
These decisions require a wide body of knowledge of the beneficial
and limiting attributes of the various project delivery methods, and
an ability to select the most appropriate methods for the situational
aspects of the particular projects in question. The results of this
research have provided conclusive project delivery method performance information that can be used to assist decision-makers in the
selection process. Thus, it is recommended that state and local district policies be aligned to allow for the widest possible selection
of project delivery methods for the construction of public school
projects, enabling those in decision-making capacities to make the
proper choices to benefit the public.
Future Research
Future research related to public school construction should include
the value analysis of high-performance designs, materials, and
equipment suitable for energy-efficient and sustainable building approaches. Additionally, in light of the many variables that drive
05016009-9
J. Constr. Eng. Manage., 2016, 142(10): 05016009
J. Constr. Eng. Manage.
cost, schedule, quality, and claims issues during the construction of
public schools, the researchers suggest a more in-depth study
through use of a direct comparison case study approach, to limit the
influences of outside effects.
Research is forthcoming regarding the previously discussed
influence that incorrect owner perceptions and unrealistic expectations combined with restrictive district policies are having on
project delivery method selections.
Downloaded from ascelibrary.org by Charlie Roberts on 11/17/16. Copyright ASCE. For personal use only; all rights reserved.
References
Abramson, P. (2012). “The 2012 school construction report.” CR1-16,
School Planning and Management, Chatsworth, CA.
Abramson, P. (2013). “The 2013 school construction report.” CR1-16,
School Planning and Management, Chatsworth, CA.
Addis, B. (2007). Building: 3000 years of design engineering and
construction, Phaidon Press Limited, London.
Akintoye, A. S., and MacLeod, M. J. (1997). “Risk analysis and management in construction.” Int. J. Project Manage., 15(1), 31–38.
Atkinson, R. (1999). “Project management: Cost, time, and quality, two
best guesses and a phenomenon, it’s time to accept other success
criteria.” Int. J. Project Manage., 17(6), 337–342.
Bennett, J., Pothecary, E., and Robinson, G. (1996). “Designing and building a world-class industry.” Centre for Strategic Studies in Construction,
Univ. of Reading, Reading, U.K.
Bullen, C. V., and Rockart, J. F. (1981). “A primer on critical success
factors.” Sloan School of Management, Massachusetts Institute of
Technology, Cambridge, MA.
Carolinas AGC. (2009). Conf. on Alternative Project Delivery Methods,
Charlotte, NC.
Chan, A. P. C., and Chan, A. P. L. (2004). “Key performance indicators
for measuring construction success.” Benchmarking: An Int. J., 11(2),
203–221.
Cox, T., Kenig, M., Allison, M., Kelley, S. W., and Stark, M.. (2011).
“Primer on project delivery.” American Institute of Architects (AIA),
Washington, DC.
D’Agostino, B., and Bridgers, M. (2010). Eleventh annual survey of
owners: Rising from the ashes of recent economic woes, FMI/CMAA,
Raleigh, NC.
Demkin, J. A., and AIA (American Institute of Architects). (2008). The
architect’s handbook of professional practice, Wiley, Hoboken, NJ.
Fitchen, J. (1986). Building construction before mechanization, MIT Press,
Cambridge, U.K.
Forester, J. (1989). Planning in the face of power, University of California
Press, Los Angeles.
Ghavamifar, K., and Touran, A. T. (2008). “Alternative project delivery
systems: Applications and legal limits in transportation projects.”
J. Prof. Issues Eng. Educ. Pract., 10.1061/(ASCE)1052-3928(2008)
134:1(106), 106–111.
Gordon, C. (1994). “Choosing appropriate construction contracting
method.” J. Constr. Eng., 10.1061/(ASCE)0733-9364(1994)120:1(196),
196–210.
© ASCE
Heady, E. J. (2012). Construction Law; the History is Ancient! Construction Connection News Letter, 〈http://www.constructionconnection.com/
blog/tag/history-of-construction-law/〉 (Jun. 26, 2012).
Kangari, R. (1995). “Risk management perceptions and trends of U.S. construction.” J. Constr. Eng. Manage., 10.1061/(ASCE)0733-9364(1995)
121:4(422), 422–429.
Kenig, M. E. (2011). Project delivery systems for construction, Associated
General Contractors of Americas, Arlington, VA.
Konchar, M. (1997). “A comparison of united states project delivery systems.” Technical Rep. No. 38, Pennsylvania State Univ., University
Park, PA.
Konchar, M., and Sanvido, V. (1998). “Comparison of U.S. project delivery
systems.” J. Constr. Eng. Manage., 10.1061/(ASCE)0733-9364(1998)
124:6(435), 435–444.
McNichol, E., Oliff, P., and Johnson, N. (2011). States continue to feel recession’s impact, Center on Budget and Policy Priorities, Washington,
DC.
National Research Council. (2009). Advancing the competitiveness and
efficiency of the U.S. construction industry, National Academies Press,
Washington, DC.
O’Connor, P. J. (2009). “Integrated project delivery: Collaboration through
new contract forms.” Faegre & Benson LLP, Minneapolis.
Oliff, P., Mai, C., and Palacios, V. (2012). States continue to feel recession’s
impact, Center on Budget and Policy Priorities, Washington, DC.
Rockart, J. F. (1979). “Chief executives define their own data needs.”
Harvard Bus. Rev., 57(2), 81–93.
Rojas, E. M., and Kell, I. (2008). “Comparative analysis of project delivery
systems cost performance in Pacific Northwest public schools.” J.
Constr. Eng. Manage., 10.1061/(ASCE)0733-9364(2008)134:6(387),
387–397.
RS Means. (2013). RS means construction cost data, 71st Ed., Construction
Publishers & Consultants, Norwell, MA.
Sanvido, V., and Konchar, M. (1999). “Selecting project delivery systems:
Comparing design-bid-build, design-build, and construction management at risk.” Project Delivery Institute, State College, PA.
Saporita, R. (2006). Managing risk in design and construction projects,
ASME Press, New York.
SAS/STAT version 9.3 [Computer software]. SAS Institute, Cary, NC.
U.S. Census Bureau. (2011). “Capital spending report 2011.pdf (application/pdf object).” Washington, DC.
U.S. DOT (U.S. Department of Transportation). (2006). “Design-build
effectiveness study.” Washington, DC.
Vincent, J. M., and McKoy, D. L. (2008). The complex and multi-faceted
nature of school construction costs: Factors affecting California,
American Institute of Architects, Sacramento, CA.
Williams, G. H. (2003). “An evaluation of public construction contracting methods for the public building sector in Oregon using data
envelopment analysis.” Ph.D. dissertation, Portland State Univ.,
Portland, OR.
Zaghloul, R., and Hartman, F. (2003). “Construction contracts: The cost of
mistrust.” Int. J. Project Manage., 21(6), 419–424.
05016009-10
J. Constr. Eng. Manage., 2016, 142(10): 05016009
J. Constr. Eng. Manage.