Public Workshttp://pwm.sagepub.

com/
Management & Policy

Asset Management Data Practices for Supporting Project Selection Decisions

Aristeidis Pantelias, Gerardo W. Flintsch, James W. Bryant, Jr and Chen Chen
Public Works Management Policy 2009 13: 239 originally published online 4 December 2008
DOI: 10.1177/1087724X08327574
The online version of this article can be found at:
http://pwm.sagepub.com/content/13/3/239

Published by:
http://www.sagepublications.com

Additional services and information for Public Works Management & Policy can be found at:
Email Alerts: http://pwm.sagepub.com/cgi/alerts
Subscriptions: http://pwm.sagepub.com/subscriptions
Reprints: http://www.sagepub.com/journalsReprints.nav
Permissions: http://www.sagepub.com/journalsPermissions.nav
Citations: http://pwm.sagepub.com/content/13/3/239.refs.html

>> Version of Record - Feb 17, 2009

OnlineFirst Version of Record - Dec 4, 2008
What is This?

Downloaded from pwm.sagepub.com at National School of Political on May 20, 2014

Asset Management Data Practices for

Supporting Project Selection Decisions

Public Works Management &

Policy
Volume 13 Number 3
January 2009 239-252
2009 SAGE Publications
10.1177/1087724X08327574
http://pwmp.sagepub.com
hosted at
http://online.sagepub.com

Aristeidis Pantelias
The University of Texas at Austin

Gerardo W. Flintsch
Virginia Polytechnic Institute and State University, Blacksburg

James W. Bryant Jr.

Transportation Research Board, Washington, DC

Chen Chen
Asian Development Bank, Pacific Department, Manila, Philippines
Transportation agencies engage in extensive data collection activities to support their decision processes at various levels,
but not all data collected supply useful information. This article summarizes research aimed at formally identifying links
between data collection and the supported decision processes, particularly at the level of project selection. The aim was to
help transportation agencies optimize their data collection and cut down data collection and management costs. The methodology included a comprehensive literature review that collected information from various academic and industry sources
around the world and the development of a Web survey that was e-mailed to specific expert individuals within the 50 U.S.
Departments of Transportation and Puerto Rico. The responses obtained from the Web survey were analyzed statistically
and combined with the additional resources to extract conclusions about the current state of the practice and develop data
collection recommendations in the form of a proposed stepwise framework.

Keywords: asset management; data collection; decision processes; decision levels; project selection; Web survey

Background
Transportation asset management is a strategic
approach to the optimal allocation of resources for the
management, operation, maintenance, and preservation
of transportation infrastructure (Federal Highway
Administration [FHWA], 1999). It combines engineering and economic principles with sound business practices to support decision making at the strategic,
network, and project levels, and its scope spans beyond
physical assets to encompass all transportation assets
within a state or local transportation agency.
One of the key aspects of the development of asset
management is data collection. The way in which transportation agencies collect, store, and analyze data has
evolved along with advances in technology, such as
mobile computing (e.g., handheld computers, laptops,
tablet notebooks, etc.), sensing (e.g., laser and digital cameras), and spatial technologies (e.g., global positioning
systems, geographic information systems, and spatially

enabled database management systems). These technologies have enhanced the data collection and integration
procedures necessary to support the comprehensive analyses and evaluation processes needed for asset management
(Flintsch, Dymond, & Collura, 2004).
The concept of asset management has been defined by
various governmental organizations, transportation agencies, and industry groups Danylo & Lemer, 1998;
Transportation Association of Canada, 1999; Organization
for Economic Cooperation & Development, 2002;
National Cooperative Highway Research Program
[NCHRP], 2002). The FHWA defines asset management
as follows:

Authors Note: Please address correspondence concerning this article to Aristeidis Pantelias, The University of Texas at Austin, 1
University StationC1761, Austin, TX 78712-1076; e-mail: a.pantelias@mail.utexas.edu.

239
Downloaded from pwm.sagepub.com at National School of Political on May 20, 2014

240

Public Works Management & Policy

[Asset management is] a systematic approach of maintaining, upgrading, and operating physical assets costeffectively. It combines engineering principles with
sound business practices and economic theory, and it
provides tools to facilitate a more organized, logical
approach to decision-making. Thus, asset management
provides a framework for handling both short- and longrange planning. (FHWA, 1999)

In many cases, however, the data collection activities

have not been designed specifically to support the decision processes inherent in asset management. As a result,
the use of the aforementioned technologies has led agencies to collect large amounts of data and create vast databases that have not always been useful or necessary for
supporting decision processes.

Objective
To support asset management, agencies must collect,
store, manage, and analyze large amounts of data in an
effective and efficient manner. Agencies have strongly
emphasized collecting and integrating data, and significant effort has been directed into linking the data collection to their decision-making processes. By focusing on
the use of the data and the nature and needs of the decision levels and processes to be supported, transportation
agencies can define what assets and which data about
these assets are more important for decision making and
tailor their data collection accordingly.
This article presents a study of the state of the practice
on how state departments of transportation (DOTs) are
linking their data collection policies, standards, and
practices to their asset management decision-making
processes. The study focused primarily on project selection, the decision-making level that functions as an intermediate stage between high-level strategic and low-level
project-specific decisions. Projects under consideration
include capital and maintenance works of physical transportation assets with a specific focus on highway infrastructure (i.e., pavements, bridges, culverts, etc.).

Methodology
The investigation started with a comprehensive literature review to retrieve related experience information
from academia and industry sources throughout the
world. Several reports have documented current and past
practices in the United States and Canada as well as in
Europe and Australia. The literature review summarizes
the state of the art in asset management, decision making,
and data collection.

To complement the literature review, a Web-survey

questionnaire was developed to capture the current level
of asset management endorsement and implementation
within state DOTs as well as specific aspects of their data
collection practices and their relationship with the project selection level of decision making. A link to the survey was distributed to specific DOT officials in all 50
states and Puerto Rico, and the responses were stored
and subsequently analyzed.
The information obtained was ultimately used to
extract conclusions about the current state of the practice
of U.S. DOTs in asset management data collection and
related decision-making activities, primarily for project
selection. It was also used to develop recommendations
for effective and efficient data collection by means of a
stepwise framework. The proposed framework aims to
help transportation agencies tailor their data collection
activities according to their real decision-making needs
and, in the same way, contribute both to the reduction of
data collection costs and to a more effective and efficient
implementation of asset management.

Literature Review
Asset Management, Data Collection,
and Decision Levels
The emerging field of asset management has been
extensively researched and many contributions have
been made to date by transportation agencies and other
related stakeholders in the United States and around
the world. In particular, the Transportation Asset
Management Guide by the American Association of
State Highway and Transportation Officials (AASHTO)
represents a milestone domestic reference (National
Cooperative Highway Research Program (NCHRP,
2002). This guide was prepared to assist state DOTs in
tailoring a generic asset management framework to their
individual needs and characteristics. It was developed in
the context of the NCHRP and based on up-to-date experience and research findings. This guide describes asset
management as a strategic approach to managing infrastructure assets and identifies two major clusters of decision making: resource allocation and use. The process
described is oriented to the actual implementation of the
asset management concepts and methodologies. Other
important recent contributions include NCHRP (2005),
NCHRP (2006), and FHWA International Technology
Exchange Program (2005); additional details on supporting literature can also be found in Pantelias (2005).
Good asset management implies a systematic integrated approach to project selection, analysis of trade-offs,

Downloaded from pwm.sagepub.com at National School of Political on May 20, 2014

Pantelias et al. / Data Collection and Project Selection Decisions 241

resource optimization, programming, and budgeting. This

form of management relies on accurate asset inventory,
condition, and system performance information; considers
the entire life cycle cost of the asset; and combines engineering principles with economic methods, thus, seeking
economic efficiency and cost-effectiveness. AASHTO
(NCHRP, 2002) presents a framework for transportation
asset management as a philosophy, process, and resource
allocation and utilization process. The framework includes
the functions for policy-driven planning based on system
performance; economic, social, and environmental objectives; integrated analysis tools to evaluate trade-off among
asset classes, performance goals, and types of investments; decisions on allocation of resources and investments; implementation of the programs through different
executing modes; and system monitoring on the system
performance evaluation to determine the degree of accomplishment of the defined goals (NCHRP, 2002).
According to FHWA (1999), an asset management system should be customer focused, mission driven, integrated (system-oriented), long term in outlook, flexible,
accessible, user-friendly, and should have explicit goals
for asset performance. This decision support system has to
support trade-offs considering costs, benefits, and other
impacts of the alternative potential investments. This
requires appropriate information gathering and processing. Thus, one of the key building blocks of any automated
asset management application or integrated asset management system is a comprehensive inventory of the highway
infrastructure assets and their respective condition. These
systems may be designed to manage one type of transportation asset (e.g., pavements or bridges) or several
types of assets in a comprehensive and integrated manner.
In terms of implementation, most transportation agencies have basic bridge and pavement data for their transportation networks; several of them have also made a
great effort over the years to collect, store, manage, and
analyze comprehensive inventory data for their other
highway infrastructure assets (Data Integration, 2001;
Larson & Skrypczuk, 2004; Virginia Department of
Transportation, 2004). Sanford and McNeil (1998) proposed a data requirements model for bridge condition
assessment and a bottom-up process that an agency can
use to assess and improve its data collection practices.
Furthermore, there has been significant work on data
integration (Data Integration, 2001; FHWA, 2004; Hall,
2005; Vandervalk-Ostrander, Guerre, & Harrison, 2003).
Asset management requires the monitoring of the
transportation systems using performance measures that
reflect the agencys goals and objectives. The NCHRP
(2006) Report 551 defined performance measurement as

a way of monitoring progress toward a result or goal,

more specifically it helps to keep track and forecast the
impacts within and outside the system. Existing performance measures and analytical approaches are used to
evaluate the effects of highway investment on economic
productivity, user benefits (that include reductions in
fatalities and serious injuries), environmental impacts,
and societal benefits, among other factors. The report
identified numerous performance measures and organized them into the following 10 categories: preservation, accessibility, mobility, operations, maintenance,
safety, environmental impacts, economic development,
social impacts, security, and delivery (FHWA, 2005).
Furthermore, in every managerial process, there exists
an internal hierarchy of decision-making levels. Higher
level decisions require a higher level of aggregation of the
supporting information and produce decisions with a
wider scope than those at the lower levels. Transportation
asset management is no exception. There are various
decision-making levels that represent different perspectives on the business functions of the agencies, ranging
from very specific, detailed, project-oriented views to
generalized, comprehensive, and strategic ones. Although
asset management is mostly perceived as a strategic-level
tool, it nevertheless affectsand can be equally successful inlower levels of decision making within a transportation agency.
The decision levels pertaining to asset management
can be broken down into strategic, network, and project
levels (Haas, Hudson, & Zaniewski, 1994). Strategic
decision making pertains to strategic resource allocation
and utilization decisions concerning all types of assets
and systems within the constructed environment, one of
them being the transportation sector. Within transportation, it may consider all different modes and all assets
pertaining to these modes. The network decision-making
level pertains to determining the overall agency-wide
maintenance, rehabilitation, construction strategies, and
work programs. This decision level considers systemwide budget allocation and transportation planning
decisions, but its scope is narrower than the strategic
levels. Project-level decision making and analysis pertains to the specific, modewise, assetwise, and geographically determined projects. It addresses the design of the
projects included in the overall work plan needed to meet
the agencys performance measures. It is also called field
level or operational level and refers to how the actual
work is going to be done.
The network level is itself often broken down into program and project selection levels (Haas, Hudson &
Zaniewski, 1994). The program decision-making level is

Downloaded from pwm.sagepub.com at National School of Political on May 20, 2014

242

Public Works Management & Policy

concerned with the overall, networkwide programming

of actions and allocations. It is involved in policy decisions, and the aim is the systemwide optimization of
funds allocated to rehabilitation, maintenance, or new
construction of infrastructure assets. The project selection level is concerned with decisions on funding for projects or groups of projects. This level generates decisions
at a higher level of aggregation than the project level, but
it requires more detailed information than the program
and network levels. It serves as a link between the network level and the subsequent project level of analysis.
As explained previously, the different levels of decision making have different roles: Higher levels are
mostly concerned with overall budget allocations and
system utilization, whereas lower levels tend to focus
more on the administration, funding, and engineering of
specific functions and processes. In addition, decision
makers have different backgrounds and interests. As a
result, the decisions at each level are different in scope
and require different aggregation of information, affecting directly the corresponding detail and quantity of the
data to be collected. Higher levels require more generalized information, whereas lower ones require more
detail. The data needed to support decision making in
each of the levels are of paramount importance to the
success of an asset management process and directly
determine the corresponding data collection costs. This
article focuses primarily on the data needs and corresponding data collection practices of the project selection level of decision making.

Decision Making at the Project Selection Level

Project selection is a level of decision making that
entails the evaluation of the attributes of different candidate projects for the purpose of funding and implementation. These projects, as mentioned before, include capital
and maintenance works for different classes of physical
assets. As a decision-making level, project selection functions as a bridge between high-level network decisions
and site-specific, detailed project-level decision making.
The project selection analysis is based on information
that is aggregate enough to show the big picture of the
competing projects and, therefore, identify and assess
their usefulness and overall impact. However, it is also
detailed enough to capture the individuality of each project, provide accurate cost estimates, and identify implementation implications for the agencies and the users.
Within the framework of asset management, the
nature of project selection is unique because in many
cases the candidate projects concern different assets and
treatments. As an example, trade-offs may consider the

rehabilitation of an existing flexible pavement through

milling and repaving versus the maintenance of a concrete pavement through crack sealing. Trade-offs may
also consider the maintenance of the roadside drainage
system of a particular segment of a highway versus the
rehabilitation of the bridges concrete deck or the
replacement of its steel railings.
According to Hudson, Haas, and Uddin (1997) project selection follows the overall network programming
decision level regarding the general funds that are going
to be allocated in different types of agency works. After
the agency has decided on the amount of funds to be
spent in maintenance, rehabilitation, or new construction (or reconstruction), the candidate projects that fall
into each of these work programs need to be determined.
The selection of the different projects or different
groups of projects to be included in a work program is
heavily constrained by available budgets and usually
resorts to some type of prioritization models. These
models usually use optimization, near-optimization, or
other ranking techniques to lead to results that can be
used by the transportation officials to support decision
making (Haas et al., 1994). Furthermore, the type of
cross-asset analysis performed at the project selection
level represents a level of decision making of its own
and is not to be confused by similar analyses that can
take place at the program level.
Worldwide practice in the area of project selection has
shown that for the analysis to be as comprehensive and
as accurate as possible, the effects of economic and timing parameters should also be considered. As a result,
recent research in this area has focused on proposing
competent models that include economic analysis and
multiyear prioritization algorithms among others, and
some of these models also take into account the effects
of risk and uncertainty in their final outcome (Li &
Sinha, 2004). The most commonly used economic analysis techniques are discussed in more detail elsewhere in
Haas, Hudson, and Zaniewski (1994), Haas, Tighe, and
Cowe Falls (2004), and Flintsch and Kuttesch (2004).
Because most agencies carry out project selection by
using specific quantitative and qualitative tools similar to
the ones described above, it is reasonable to assume that
the data needs of this particular level of decision making
should include the type and amount of data that form the
inputs of these models and techniques. In other words,
project selection data needs should be focused on the
particular inputs that the project evaluation models
require. As different agencies use different models and
techniques, the particular data needed for each agency
are also bound to be different.

Downloaded from pwm.sagepub.com at National School of Political on May 20, 2014

Pantelias et al. / Data Collection and Project Selection Decisions 243

Data Characteristics and Properties

Another important issue of data collection relates to the
particular attributes and characteristics that the collected
data should possess to be useful in supporting decision
making. Regardless of the projected use of the collected
data, it is of paramount importance that they exhibit the
following characteristics (Deighton, 1991, chap. 3):
Integrity: Whenever two data elements represent
the same piece of information, they should be equal
(in value or meaning).
Accuracy: The data values should represent
as closely as possible the considered piece of
information.
Validity: The given data values should be correct in
terms of their possible and potential ranges of
values.
Security: Sensitive, confidential, and important
data are protected by restricting access to them and
by properly ensuring systematic and frequent backing-up in other storage media.
Furthermore, the timeliness of the data is also important, so that the data reflect the current characteristics and
condition of the assets. It is also recommended that the
data elements be rigorously defined in a data dictionary
and thatin the most ideal of all casesthese definitions
be common between all agencies and parties involved in
this area of practice (Deighton, 1991, chap. 3).
In addition, the Western European Road Directors
(WERD, 2003) guidelines highlighted the importance of
the following criteria when selecting data required by an
agency/organization:
Relevance: Every data item collected and stored
should support an explicitly defined decision need.
Appropriateness: The amount of collected and
stored data and the frequency of their update should
be based on the needs and resources of the agency
or organization.
Reliability: The data should exhibit the required
accuracy, spatial coverage, completeness, and
currency.
Affordability: The collected data should be in
accordance with the agencys financial and staff
resources.
Consequently, agencies planning to engage in data collection should take into account the specification of the data
to be collected, the frequency of collection, the accuracy
and quality that they should exhibit, and their completeness

and currency (WERD, 2003). As a general recommendation, it is noted that the accuracy, quality, and timeliness of
the data should be decided based on the cost of the data collection and the value and benefit associated with the data in
question: Data should only be collected if the benefits that
they provide outweigh the cost of their collection and maintenance (WERD, 2003). Data collection costs can and
should be minimized by collecting only the needed data
when they are needed. The data collection activities and
methods used should be based on the requirements of the
decision processes to be supported and should produce
results that match the levels of accuracy, precision, and resolution required by these decision processes (Smith &
Lytton, 1992).

Web-Based Transportation Asset

Management Survey
To complement the literature review and to capture
the corresponding state of the practice in the United
States and Puerto Rico, a two-part Web-based survey
was developed and posted on the Internet. The survey
was sent to 103 transportation officials from all 50 states
and Puerto Rico, and responses were stored in a specifically designed database and subsequently processed and
analyzed.
The first part of the survey, entitled General Agency
Information on Asset Management, contained questions
on the following:
asset management endorsement and implementation by the state DOTs;
existing and/or planned management systems for
specific asset types;
existing levels of decision making within state
DOTs;
identification and rating (based on their perceived
importance) of existing asset management decision
processes and functions within state DOTs; and
identification and rating of existing criteria used by
the agencies for project selection.
The second part of the questionnaire, entitled
Roadway Asset Management, required more specialized and detailed input on current agency practices on
data collection and project selection, and it contained
questions on the following:
data management, collection methods, and integration;
rationale behind existing and/or future planned data
collection;

Downloaded from pwm.sagepub.com at National School of Political on May 20, 2014

244

Public Works Management & Policy

evaluation of roadway asset data used for project

selection; and
identification of formally documented links
between data collection and project selection or
asset management decision processes in general.

Therefore, the responses used in the analysis are a

combination of the responses provided by the different officials.
2. In the remaining cases, in which the survey questions asked for individual opinions, all 48 answers
from all responding officials were considered
valid and were used in the analysis.

The contents of the survey were refined several times

for suitability of content, wording of the questions, and
overall format. Comments and input were requested and
received from the Statistics and Survey Department of
Virginia Tech as well as from the AASHTO expert task
group supervising the project. The authors will be happy
to provide copies of the complete survey forms that were
used to create the Web-based questionnaire on request
from readers.

Answers to essay questions were not considered in

the statistical analysis but were rather used as a guide
for the resolution of discrepancies and also as a compass for the overall status of the responding state in
relation to the researched topics.
In the following sections, the results obtained from the
survey responses are summarized and discussed.

Survey Analysis

Asset Management System Implementation

A total of 48 responses from 40 different states were

received. Based on the initially targeted number of states and
respondents, a response percentage of 78% was achieved in
terms of individual states and 47% in terms of individual
respondents. The respondents were the main asset management contacts (e.g., director of the asset management division) for the state DOTs as identified by FHWA.
For seven states, more than one transportation official
responded to the survey. Furthermore, whereas some of
the questions on the survey were state specific and
required only one valid answer per responding state,
others requested the personal opinion of the responding
state transportation officials. Because of this difference,
two approaches were followed in processing the
responses of the states that provided more than one
response:

The responses to the first question concerning the

implementation stage of an asset management system
revealed that a significant portion of the responding
states (24 out of 40) are still in the planning phase. Only
approximately one quarter (11) of the respondents indicated that they have already implemented an asset management system.
The responses also revealed that most of the responding
states have been using stand-alone management systems
for specific asset types, the most predominant among
them being pavement (39), bridge (39), and maintenance
(34) management systems. Other systems include safety
(SMS), congestion (CMS), public transportation (PTMS),
and intermodal transportation (ITMS) management.
However, for most of these states, the level of integration of these individual systems within an overall asset
management framework is still in the planning phase.
Pavement and bridge management systems seem to be
one step ahead of the remaining systems in terms of this
integration.

1. In the first case, the various answers within the

same state were compared, and discrepancies were
resolved, so that only one answer that was as complete as possible was kept, based on the following
criteria:
Priority was given to the most complete
response; for example, in the case in which one
transportation official reported that the state
agency possessed two management systems for
specific asset classes and another official
reported the possession of three, the final
response contained all three individual management systems from the second response.
Priority was given to the responses of transportation officials whose area of expertise
most closely coincided with that of the surveys
questions and input fields.

Decision Levels and Processes

When asked to report on their defined decision-making levels, most of the responding transportation agencies indicated that they have explicitly defined
decision-making levels (Figure 1). These decision levels
in general coincide with the ones found in the literature.
The decision-making levels most frequently identified
by survey respondents were programming and budgeting,
and project selection. This finding implies that the agencies have been focusing their attention on these intermediate levels connecting the generic strategic decisions of the
strategic level to the actual project implementation at the

Downloaded from pwm.sagepub.com at National School of Political on May 20, 2014

Pantelias et al. / Data Collection and Project Selection Decisions 245

Figure 1
Defined Decision-Making Levels

project level. It also suggests that the right transportation

officials were selected to complete the survey because
they were familiar with these levels of decision making.
Further on, the state transportation officials were
asked to rank a list of identified asset management decision processes in terms of their importance. Figure 2
summarizes the responses and shows that most of the
listed decision processes fall in the very important and
somewhat important categories.
The relative importance of the decision processes was determined by computing the average importance rating for each
decision process using a score from 1 to 4 (4 = very important
to 1 = not important at all). The resulting ranking reflects the
perceived relative priorities of the responding states. The most
important decision process reported by the agencies surveyed
was performance evaluation and monitoring, with fiscal
planning following close behind. Project selection, which is the
main interest of this investigation, ranked third together with

resource allocations, indicating the perceived significance of

this decision process among the responding transportation officials. On the other hand, the low ranking of processes such as
development of alternatives and impact analysisprocesses
important in understanding cost trade-offs and performing
investment analysescannot be left unnoticed. This shows
that a lot of states have not fully understood some of the basic
characteristics of asset management and demonstrates a significant gap between state of the art and state of the practice in
asset management.
Figure 3 summarizes the importance assigned by the
state transportation officials to a list of specific project
selection criteria. The variability of opinions is more
significant in this question. The criterion of available
budgets/earmarked funds stands out as most important,
followed closely by engineering parameters. The bottom
part of the figure shows the average rankings for all the
listed criteria based again on their computed relative

Downloaded from pwm.sagepub.com at National School of Political on May 20, 2014

246

Public Works Management & Policy

Figure 2
Asset Management Decision Processes and Their Relative Importance

Downloaded from pwm.sagepub.com at National School of Political on May 20, 2014

Pantelias et al. / Data Collection and Project Selection Decisions 247

Figure 3
Project Selection Criteria and Their Relative Importance

Downloaded from pwm.sagepub.com at National School of Political on May 20, 2014

248

Public Works Management & Policy

importance. An interesting finding from this question is

that public demands/user opinions rank third, showing
the increased interest of transportation agencies in public satisfaction with the selection and implementation of
projects.
Furthermore, and to the surprise of the research team,
the vast majority (80%) of the responding officials
agreed that the criteria used for project selection cannot
or should not be uniform for and consistent across all
types of roadway assets. This, however, presents a contradiction to known common practices identified in the
literature review where cross-asset comparisons are
performed through costbenefit, life cycle costing, and
various other engineeringeconomic analyses and prioritization/optimization methods. Although agency, user,
and community costs and benefits were proposed as criteria for such comparisons, none of them appeared in the
top three relative importance ranking positions denoting
that there is a lack of agreement between the literature
review and the survey results. This result, again, can be
interpreted as a lack of understanding of what good asset
management practices entail, supporting the previously
mentioned lag between the state of the art and the state
of the practice in asset management.

Data Collection Procedures

Most of the responding state agencies (75%) have already
invested time and money in developing asset management
roadway inventories and databases. The majority of the
remaining agencies reported that they were planning for them.
Most agencies have also been collecting data predominantly
for their pavements and bridges. Traffic items (e.g., attenuators, guardrail, pavement striping, pavement markings) and
roadside assets (e.g., trees, historic markers, right-of-way
fence) also made up a large portion of the collected data.
Figure 4 summarizes the data collection methods used
for the acquisition of the above data. Whereas for some
assets (e.g., drainage) the collection is reported to be taking
place mostly manually, there is a trend toward using a combination of manual and automatic methods. This is consistent with what was reported recently by Flintsch, Dymond,
and Collura (2004).

Data Collection Rationale

The officials were also asked to provide information
about their rationale behind data collection. The responses
appear to support the findings from the literature review
that most agencies still base their data collection decisions
(e.g., which data to collect) on past practices and staff

experience. However, many respondents also noted that

data collection practices have been based on data collection standards and input needs of management systems
used or other defined decision processes. This confirmed
the initial finding from the literature review that there is an
ongoing effort to link data collection to specific decision
processes.
The next question asked state officials to rate the importance of identified roadway asset data in choosing between
two competitive projects. The ratings are summarized in
Figure 5. From the responses obtained, the most important
data are the assets structural and functional conditions,
with usage of the assets following in third place.
The results indicate that responding agencies tend to
use the same kind of data to prioritize projects between
different types of transportation infrastructure assets.
Again, however, life cycle costs were not one of the top
three responses (although these costs were cited extensively in the literature review as key to project trade-off
analysis and comparisons).
Finally, the last question of the survey investigated the
level at which state transportation agencies were conscious about the existence of links between their data
collection activities and project selection. From the
responses, it was determined that several agencies have
identified (32.5%) or identified and documented in a formal way (52.5%) the existence of such links. This is an
important finding because it shows that most agencies
(85%) have been trying to rationalize their data collection according to specific decisions to be supported, at
least for the project selection level. This is also consistent with the literature that reports significant efforts to
improve data collection and management practices.

Recommended Framework for

Data Collection
The literature review confirmed that research in the
area of asset management and the related data collection
has been extensive. Little information, however, could be
found concerning specific data collection to support project selection within the framework of asset management. As this level links the overall network decisions
with the individual projects to be implemented, it consequently requires data that are simultaneously neither too
generic nor too specific. Generic data, such as networklevel performance indices, would not help in project
selection because they would ignore vital project details.
On the other hand, it is usually not cost-effective to collect very detailed (e.g., project-level) data for the project

Downloaded from pwm.sagepub.com at National School of Political on May 20, 2014

Pantelias et al. / Data Collection and Project Selection Decisions 249

Figure 4
Roadway Asset Data Collection Types and Methods

selection process, although such bottom-up approaches

can also be used in cases where such data already exist.
Furthermore, project selection has traditionally been
made between projects that belong to the same asset
class. Asset management supports the broadening of this
traditional practice by encouraging cross-asset comparisons between the candidate projects for selection. This
has obviously modified the data needs and has also created the need for the identification and use of effective
selection methodologies that can be applied equally and
in an unbiased way to all different asset classes. Such
methods were identified in the literature to some extent
but not in the survey.
The survey of current practice suggests that U.S. transportation agencies have clearly identified decision-making

levels and also have relatively uniform perceptions of the

importance of various asset management decision
processes, project selection criteria, and corresponding
asset data that could support selection between competing
projects.
A proposed framework for effective and efficient
data collection is presented in Figure 6. This framework leverages findings from the literature review and
results from the survey of state transportation officials. It can help agencies determine which data are
needed and customize their data collection practices
accordingly.
For an agency to evaluate its data collection needs for
project selection, the transportation officials should
answer the following questions:

Downloaded from pwm.sagepub.com at National School of Political on May 20, 2014

250

Public Works Management & Policy

Figure 6
Proposed Framework for Project Selection Data Collection

Downloaded from pwm.sagepub.com at National School of Political on May 20, 2014

Pantelias et al. / Data Collection and Project Selection Decisions 251

Figure 6
Proposed Framework for Project Selection Data
Collection

STEP 1

Identification of typical candidate projects and

corresponding preservation and rehabilitation treatments
for all assets under consideration

STEP 2

Identification of project evaluation/assessment tools and

their data requirements

STEP 3

Identification and/ordefinition of data needs for the

analyses/evaluations to be performed

STEP 4

Investigation of existing databases, available data, data

collection practices, and data quality assessment
procedures

STEP 5

Identification of missing data elements

STEP 6

Selection of the appropriate data collection methodology

for the missing data elements and implementation

(Feedback to
upper levels)

FEEDBACK
LOOP:
Evaluation of
assessment
tools/ data
inputs/ data
collection and
overall results

What are the various types of roadway assets in

need of preservation, rehabilitation, or renewal
actions? (Step 1)
What are the various types of preservation, rehabilitation, or renewal treatments that should be considered? (Step 1)
What are the evaluation models, techniques, and
criteria used by the agency to assess the usefulness
of the projects and prioritize/optimize their selection? (Step 2)
What are the inputs required by these models and
techniques for the various projects to be assessed?
(Step 3)
What are the data that already exist in available and
accessible databases and/or are already being collected? (Step 4)
What additional data need to be collected? (Step 5)
Once the needed data have been identified, the agencies can decide on the most appropriate method to obtain
them. This may mean either extracting them from
already existing databases (where applicable and possibly through additional processing) or engaging in additional data collection activities and correspondingly
deciding on their required level of accuracy, precision,
and resolution (Step 6).
As a final element of the proposed framework, a feedback loop should be established after the missing data have
been collected and the project selections for at least one

cycle have been performed to evaluate the effectiveness of

the models and their results and to further refine the models, data inputs, databases, and collection methods.
This framework, although generic, summarizes the
basic principles and best practices behind data collection
and can function as a starting point for transportation
agencies that wish to handle project selection in a more
systematic way. By rationalizing their data collection
needs, agencies can cut down the collection of unnecessary data and achieve significant cost reductions.
It is, however, obvious that the use of the proposed
framework would only lead to a partial optimization of
an agencys data collection activities. This framework
addresses only project selection decisions without taking
into account the needs of the other levels of decision making (e.g., network-level resource allocation or project-level
design) that might require overlapping or complementary
data and may warrant new, modified, or additional data
collection activities.
For an overall true optimization of data collection, the
data needs of all levels of decision making (including
strategic, program, and project levels in addition to the
project selection level) should be taken into consideration, and a more comprehensive framework for data collection should be established. For this purpose, the
findings of the survey concerning the most important
asset management decision processes can be used. These
decision processes can be attributed to all different decision-making levels, and therefore, the consideration of
the individual data needs of each one and their inclusion
in a generalized version of the proposed framework can
lead to a higher level of data collection efficiency.

Conclusions
Asset management has been endorsed by most state
transportation agencies as well as other major transportation
agencies throughout the world. Asset management principles have been used to manage individual classes of assets
(such as pavement and bridges) for years. The state of the
art has been steadily advancing, and significant contributions have been made by various stakeholders. However,
implementation of integrated asset management is still at its
initial stages, and there are many hurdles to overcome. In
this respect, the development of integrated roadway inventories and databases is still under way in many agencies and
so is the integration of individual management systems.
The survey results can be considered as a snapshot
of the degree to which asset management has penetrated
the industry to date and lead to some very interesting
conclusions. Transportation agencies in the United States

Downloaded from pwm.sagepub.com at National School of Political on May 20, 2014

252

Public Works Management & Policy

have explicitly defined decision-making levels and are

moving forward to a rationalization of their data collection activities. Asset management practitioners, in general, agree that project selection criteria cannot or should
not be uniform and consistent for all asset types considered. This is, however, inconsistent with known common
practices documented in the literature where costbenefit analyses or life cycle costing, among other methods,
are used toward that end and should be a topic for further
research and clarification. Past agency practices and staff
culture are still the predominant decision factors behind
data collection but have started to give way to decisions
based on data collection standards and input needs. The
structure of the funding program, which is often aligned
by asset categories, may also discourage agencies from
making cross-asset life cost analyses In the particular
area of project selection, there also seems to be a formally established relationship between the data collected
and the decisions supported.
A data collection framework for project selection is recommended to optimize the data collection activities for
project selection. Using a data collection framework, such
as the one proposed in Figure 6, can help provide clear and
logical steps toward the rationalization of the data needs
for these decisions. This framework, however, can only
partially optimize the overall agency data collection activities because it only addresses project selection decisions.
Additional effort is needed to generalize the proposed data
collection framework for an overall data collection optimization, taking into account all agency decision levels.
This research aims to help transportation agencies tailor their data collection activities according to their real
decision-making needs. In this way, it intends to contribute both to the reduction of data collection costs and
to a more effective and efficient implementation of asset
management in everyday practice.

References
Danylo, N. H., & Lemer, A. (1998). Asset management for the public
works manager: Challenges and strategies. Findings of the APWA
task force on asset management. Washington, DC: APWA.
Data Integration for Asset Management Forum and Peer Exchange.
(2001). Proceedings data integration for asset management forum
and peer exchange. Retrieved October 21, 2008, http://openlibrary
.org/b/OL3749208M
Deighton, R. (1991). An advanced course on pavement management
systems. Boston: TRB, FHWA, and AASHTO.
Federal Highway Administration. (1999). Asset management primer.
Washington, DC: Office of Asset Management, Federal Highway
Administration.
Federal Highway Administration. (2004). Data integration: The
Virginia experience (Transportation Asset Management Case
Studies FHWA-IF-04-012). Washington, DC: Author.

FHWA International Technology Exchange Program. (2005).

Transportation asset management in Australia, Canada, England,
and New Zealand (FHWA-PL-05-019). Washington, DC: Author.
Flintsch, G. W., Dymond, R., & Collura, J. (2004). Pavement management applications using geographic information systems
(National Cooperative Highway Research Project 20-5, Synthesis
Topic 34-11). Washington, DC: Transportation Research Board.
Flintsch, G. W., & Kuttesch, J. (2004, January). Application of engineering economic analysis tools for pavement management. Paper
presented at the 83rd Annual TRB Meeting, Washington, DC.
Haas, R., Hudson, W. R., & Zaniewski, J. (1994). Modern pavement
management. Malabar, FL: Krieger.
Haas, R., Tighe, S., & Cowe Falls, L. (2004, October). Generic protocol for whole-of-life cost analysis of infrastructure assets. Paper
presented at the 6th International Conference on Managing
Pavements, Queensland, Australia.
Hall, J. (2005). Enhancing the value of data programs: A peer
exchange. Retrieved October 21, 2008, from http://onlinepubs.trb
.org/onlinepubs/circulars/ec077.pdf
Hudson, W. R., Haas, R., & Uddin, W. (1997). Infrastructure management systems. New York: McGraw-Hill.
Larson, C. D., & Skrypczuk, O. (2004, January). Comprehensive data
collection supporting asset management at Virginia DOT. Paper
presented at the 83rd Annual TRB Meeting, Washington, DC.
Li, Z., & Sinha, K. C. (2004, January). A methodology for multicriteria decision making in highway asset management. Paper presented at the 83rd Annual TRB Meeting, Washington, DC.
National Cooperative Highway Research Program. (2002).
Transportation asset management guide. Retrieved October 21,
2008, from http://downloads.transportation.org/amguide.pdf
National Cooperative Highway Research Program. (2005). Reports
545: Analytical tools for asset management. Retrieved October
21, 2008, from http://onlinepubs.trb.org/onlinepubs/nchrp/
nchrp_rpt_545.pdf
National Cooperative Highway Research Program. (2006). Report
551: Performance measures and targets for transportation asset
management. Retrieved October 21, 2008, from http://onlinepubs
.trb.org/onlinepubs/nchrp/nchrp_rpt_551.pdf
Organization for Economic Co-operation and Development. (2002).
Asset management for the roads sector. Paris, France: Programme
of Co-Operation in the Field of Research on Road Transport and
Intermodal Linkages, Organization for Economic Co-operation
and Development, Directorate for Science, Technology and
Industry, Expert Group on Asset Management Systems.
Pantelias, A. (2005). Asset management data collection for supporting decision processes. Unpublished masters thesis, Virginia
Polytechnic Institute and State University, Blacksburg.
Sanford, K., & McNeil, S. (1998). A practice-oriented approach to developing a data requirements model for bridge condition assessment
(Transportation Research Record 1642). Washington, DC: TRB.
Smith, R. E., & Lytton, R. L. (1992). Graduate course in pavement
management. Washington, DC: FHWA.
Transportation Association of Canada. (1999). Primer on asset management. Ottawa, Ontario, Canada: Author.
Vandervalk-Ostrander, A., Guerre, J., & Harrison, F. (2003). Review
of data integration practices and their application to transportation asset management. FHWA-IF-03-023.
Virginia Department of Transportation. (2004). Asset management data collection guide. Richmond, VA: Virginia Department of Transportation.
Western European Road Directors. (2003). Data management for
road administrations: A best practice guide. Retrieved from
www.roaddata.org

Downloaded from pwm.sagepub.com at National School of Political on May 20, 2014