Toward a Trust-Based Construction Management

Computer-Aided Civil and Infrastructure Engineering 25 (2010) 253–268

Toward a Trust-Based Construction Management

Annie Guerriero

Public Research Centre Henri Tudor, 29 Avenue J.F. Kennedy, L-1855 Luxembourg-Kirchberg, Luxembourg;Research Centre in Architecture and Engineering, 2 rue Bastien Lepage, F-54001 Nancy, France

Sylvain Kubicki∗

Public Research Centre Henri Tudor, 29 Avenue J.F. Kennedy, L-1855 Luxembourg-Kirchberg, Luxembourg

&

Gilles Halin

Research Centre in Architecture and Engineering, Nancy, France

Abstract: The growing complexity of AEC projects isincreasing the importance of the building constructioncoordination role. Moreover, the uncertainty linked tothe environment of the construction activity makes wayfor the notion of trust. The coordinator can make use ofmultiple tools/views for accomplishing his mission. Thisresearch work suggests analyzing data coming from thesedifferent views to consolidate trust indicators informingthe coordinator about “trust in the correct progressionof the construction activity.” The approach suggestedin this article distinguishes between four aspects of theactivity determining the global trust level: task progress,actor’s performance, documents required to perform thetask, and building elements resulting from the task. Theproposal suggests introducing these trust indicators in adashboard, included in a multiview interface, thus allow-ing the coordinator to identify the tasks with a low trustlevel and to understand the nature of dysfunctions. Aprototype has been developed and integrated in a service-based IT infrastructure. Results of an experiment stageare finally discussed to validate the approach.

∗To whom correspondence should be addressed. E-mail: [email protected].

1 INTRODUCTION

The AEC sector has some specific characteristics, dis-tinguishing it from other industrial sectors. Indeed, theheterogeneity of actors’ teams, which are constituted forthe duration of the project, leads to the difficulty to cre-ate and maintain enduring relationships. The nature ofthe object to be produced is also fundamentally differ-ent. The execution of a building is in a sense the exe-cution of a prototype. Consequently, the uncertaintiesrelated to its design and production environment arenumerous (e.g., nature of the ground) (Chemillier, 2003;Jin and Doloi, 2009). Therefore, risks of dysfunction aremultiple.

The building site constitutes a particular environ-ment and it is the place of numerous potential dysfunc-tions that first can be linked to documents (e.g., prob-lems of update), then linked to interactions betweenactors (e.g., lack of trust that limits exchanges), andfinally, linked to tasks and their execution (e.g., diffi-culty of performing a construction task) (Tahon, 1997).Thus, in such a context and because of the increasingcomplexity of construction projects, coordination be-comes more and more important for limiting the impactof diverse dysfunctions appearing during the buildingconstruction.

C© 2009 Computer-Aided Civil and Infrastructure Engineering.DOI: 10.1111/j.1467-8667.2009.00628.x

254 Guerriero, Kubicki & Halin

In addition, trust is a central concept in the AEC co-operation context where the autonomy and the senseof the stakeholders’ responsibility are essential to guar-antee the quality of the production (Bobroff, 1994).Trust is important for encouraging cooperation be-tween people, to surmount risk and enable action in anenvironment characterized by numerous uncertainties(Luhmann, 1988).

Thus, this research work suggests establishing abridge between the notion of trust and the coordinationassistance tools. It proposes a new form of coordinationbased on the representation of trust and focuses on theconstruction manager’s activity.

This stakeholder is in charge of assuming the schedul-ing, the coordination, and the management of thebuilding construction activities. More precisely, duringthe execution phase, he is in charge of defining the de-tailed execution planning, as well as updating it whennecessary to maintain the global execution time. He isalso responsible for the follow-up and the monitoringof the budget. To carry out the different aspects of hismission, diverse tools are at the disposal of the construc-tion manager. Two categories can be distinguished: cur-rent and emergent tools. Among the frequently usedtools, we may cite Gantt, Pert scheduling methods, orsome tools such as word processors allowing to write thebuilding site meeting report and to synthesize the pointsof dysfunction and the decisions taken. Other moreemergent tools can also be identified, such as documentmanagement platforms, more recent scheduling meth-ods (e.g., CONSCOM, Adeli and Karim (1997); Karimand Adeli (1999b); Senouci and Adeli (2001)), 4D sim-ulation tools (Sadeghpour et al., 2004; Chau et al., 2005;Rebolj et al., 2008) putting into relation a 3D modelingof a building and its planning (Rekapalli et al., 2009; Chiet al., 2009), or performance evaluation systems (Arslanet al., 2008) permitting the evaluation of the stakehold-ers’ performance (e.g., quality of production, comple-tion of the work on time, etc.).

A major limit of these tools is that they only offer apartial vision of the cooperation context. A tool suchas a dashboard, which would synthesize data comingfrom these different views, could constitute a good de-cision support system and support a more global per-ception of the cooperation context. So the aim of thisresearch work is to propose a dashboard supporting atrust-based management. This dashboard relies on therepresentation of trust for assisting the coordination ofbuilding construction activity. If there is a great amountof research considering the notion of trust at the cen-ter of e-commerce applications (e.g., eBay), there areonly few works concerning trust in the AEC domain.In addition, these research works only consider trustin actors (e.g., Zolin et al., 2000; Ekstrom, 2004; Uden

and Naaranoja, 2007). The approach of trust suggestedin this research work is slightly different because it re-lies on a more global approach of trust in the “correctprogression of the activity” (i.e., trust in each aspect ofthe activity: progress of the activity, actor in charge ofperforming the activity, building element resulting fromthe activity, and document required for performing theactivity).

In this article, Section 2 addresses the theoretical as-pects linked to the notion of trust applied to collectiveactivities. Then in Section 3, the article focuses on thenotion of context linked to the cooperative activity. Fur-thermore, the trust criteria within the framework of abuilding construction activity are identified in Section 4.These theoretical approaches enable the proposal of adashboard tool based on trust. Its service-based IT im-plementation is detailed in Section 5. Finally, the val-idation phase with AEC practitioners is described inSection 6.

2 STUDY OF TRUST WITHIN THE AEC SECTOR

This section concerns trust within the framework of acooperative activity performed during the building ex-ecution stage. Trust constitutes indeed an essential el-ement for cooperation because the achievement of acommon objective cannot be serenely envisaged with-out trust in a context where there are interdependencesbetween stakeholders.

2.1 Definition and essential concepts

If scholars do not converge on a unique definition oftrust, it is merely because it can take diverse forms(Rousseau et al., 1998). The notion of trust is often as-sociated with positive expectations about the behavioror intentions of another person (Deutsch, 1962). Litera-ture describes trust as a device enabling to overstep thecomplexity of the environment (Luhmann, 1988) andstates the limited character of trust in a given context(Hardin, 2000; Chang et al., 2006).

The examination of trust allows highlighting its di-verse concepts. First, the trust relationship is a relation-ship that exists between a “trustor” (person who trusts)and a “trustee” (in whom/in which the trust is reposed).In literature, the trustee can take diverse forms: an ac-tor, an organization, or an artifact (e.g., software, web-site, etc.) (Sutcliffe, 2006). Then, the second importantnotion is related to the situation in which the trustortrusts a trustee. Indeed, we trust somebody in the frame-work of a particular situation, and for a particular activ-ity. We suggest retaining that trust is established in threesequences (see Figure 1):

Toward a trust-based construction management 255

Fig. 1. Concepts of trust.

• The perception of the context is the first one. It allowsthe trustor to determine if he thinks the trustee to betrustworthy for performing the activity under consid-eration. This stage relies on the available knowledgeabout the trustee (e.g., reputation, competence, etc.),and about the situation (e.g., risk).

• The decision of trust is the second sequence where thetrustor decides to act in trust according to the knowl-edge that he has at his disposal.

• The action is the last one. The trustor delegates theobject of trust. Therefore, he depends on the trusteeand becomes vulnerable (Mayer et al., 1995). In thiscontext, the trustor is perfectly conscious that hetakes a risk but trust allows him to surmount it.

2.2 Sources of trust

The examination of Zucker and Kramer’s works has al-lowed us to identify principal sources of trust (Zucker,1986; Kramer, 1999). The aspects related to psychologyare voluntarily excluded because they are less relevantin the framework of this analysis. Therefore, sources oftrust can be distinguished between:

• Trust based on characteristics. This trust is based oninternal characteristics of the individual, such as cul-ture and the group he belongs to. If we considera building construction activity, this form of trustbetween actors coming from a same category is pre-dominant. Architects, engineers, or contractors con-stitute three groups well marked inside which trustnaturally exists.

• Trust coming from a third party. This type of trustrelies on the notion of reputation. If we consider abuilding construction activity, teams are short-lived,and consequently, reputation plays an important roleand determines an “a priori trust” based on ex-changes with third parties.

• Trust coming from previous experience. This trustis based on past successful references. If we con-sider a building construction activity, experiencecoming from former AEC projects strongly affectstrust.

• Trust coming from the role. This trust corresponds toa trust relative to the performance of an actor ac-cording to the role he plays within an organization.If we consider a building construction activity, rolesare clearly defined as well as responsibilities. Theseroles generate precise expectations concerning com-petences and know-how and therefore affect trustrelationships.

• Trust based on rules. This type of trust is based oncontractual mechanisms, rules, certification organi-zation, or norms. In the AEC sector, a large num-ber of certifications progressively appear (e.g., certi-fications related to competence of actors, quality ofthe building elements, etc.), standard contracts (e.g.,contracts describing the mission of the stakehold-ers), norms (e.g., norms concerning the execution ofbuilding elements). These certifications strongly af-fect trust within an organization.

3 COOPERATIVE ACTIVITY “CONTEXTS”

Up to now, the theoretical framework of trust withinthe AEC activities, which this research work deals with,has been determined. The context appears as very im-portant within the trust mechanisms. We believe thatthe interest of new generations of IT-based tools is torepresent and take into account the context of theirusers. We will now focus on the notion of contextand on its different understandings within IT-supportedcooperation.


Fig. 2. Three contexts for action.

3.1 Three contexts of collective action

The study of the cooperative activity allows highlight-ing three different types of contexts: the cooperation,the actor’s, and the user’s contexts (Halin and Kubicki,2008) (see Figure 2).

1. Cooperation context describes the collective di-mension of the activity. The generic elementsconstituting every cooperation context are thefollowing:• Actor. This concept refers to a human resource

included within an organization and taking partin the execution of the activity.

• Activity. It is decomposed and structured. Itsexecution constitutes a common goal for theactors.

• Building element. This concept also results fromthe activity. The execution of the building ele-ment concretizes the common goal of the actors.

• Document. This concept refers to “definitive”or “intermediate” results of the activity. Doc-uments are required to perform constructiontasks.

2. The actor’s context refers to the knowledge manip-ulated by the actor and to the cognitive processes,which he carries out in preparation for his indi-vidual action. Knowledge mobilization and treat-ment mechanisms are intimately linked to the ac-tor’s business skills and to his point of view on thecooperative activity.

3. The user’s context is situated between the coopera-tion and the actor’s contexts. It considers the actoras a user of computing tools. Such tools consist ofdevices enabling the perception of a cooperativeactivity context. Taking this context into accountis essential when we try to design activity supporttools. Indeed, this context allows us to consider the

tool as a mediator between the actor and the activ-ity. It highlights the fact that a tool must not onlytake into account the collective activity, but also toadapt itself to its user.

3.2 Trust in the actor’s context

The actor’s context refers to the cognitive mechanisms,with which the actor “proceduralizes” (in the sense ofGachet and Brezillon (2005)) a part of his knowledge(related to the cooperation context), to adjust his ac-tion. However, it is important to insist on the uncer-tainty that exists in a collective activity and more pre-cisely, on the achievement of the expected results. Thisuncertainty results from the individual’s limited ratio-nality (Simon, 1959) and from the unpredictability ofthe environment, which makes that every action can-not be envisaged independently of its context (Such-man, 1987). Seeing that the action is completed withinan uncertain context, two aspects seem fundamental:trust and risk. The question of trust is essential when theaction in a cooperation context is under consideration:trust in the correct progression of the collective activ-ity, trust in the achievement of the expected results, andtrust in human resources. Concretely, the question oftrust in each aspect of the cooperative activity is central(see Section 3.1) (see Figure 3). The assessment of thesetypes of trust seems essential to adjust action to the co-operation context. Let us consider, for example, an ac-tivity with a weak trust level. In this situation, an actionperformed on the cooperation context could be a con-trol action (e.g., time, resource, or documents control),which could contribute to reduce risk and consequentlyto increase trust.

The following synthesis characterizes the fundamen-tal elements of this approach:• Cooperation context. It is an information source me-

diated by tools and perceived by the actor.


Fig. 3. Trust and contexts of action.

• Actor’s context. It refers to the actor’s knowledge.Only a part of this knowledge is “proceduralized” inpreparation for action. Moreover, it allows the actorto determine the trust in the correct progression ofthe activity and consequently to adjust his action tothe cooperation context.

• Trust. It refers to trust in each aspect of the cooper-ation context (i.e., trust in the progress of the task,in the documents linked to this task, in the buildingelement resulting from this task, and in the actor’sperformance).

• User’s context. It is mediated by the IT tool and allowsthe user to obtain a contextual visualization adaptedto him.

• Perception of the cooperation context. Tools and ac-tors’ business skills guide the perception of the coop-eration context. It is related to the capacities of visu-alization techniques implemented in the IT tools usedby the actor.

• Adapted action to the cooperation context. It refersto an action adapted to its context and guided bythe trust perception in each aspect of the coopera-tion context (progress of the task, actors, building el-ements, and documents).

• Experience feedback. It is essential in the concept oftrust, because trust is built on the basis of previousexperience.

4 MEASURE OF TRUST INDICATORSFOR ASSISTING COORDINATION IN THE

AEC PROJECTS

As seen before, this particular approach of trust withinthe cooperative activity leads to four types of specific

trust (i.e., task progress, actor, document, and buildingelement). This section aims to identify the criteria andthe method for measuring the value of trust.

4.1 Identification of the trust criteria

The stage of identification of the trust criteria consistsin exploring the elements that affect the correct pro-gression of the building construction activity (i.e., incoherence with the expectations). These trust criteriathen permit calculating trust indicators which are dis-played in a dashboard to inform the construction man-ager about the potential dysfunctions on the buildingsite.

Even if there is proximity between the notions of trustand risk, the research presented in this article differsfrom other studies related to the risks on the buildingsite (Klemetti, 2006; Zou et al., 2007; Boone, 2007) be-cause our objective is slightly different. In this work, theaim is making use of the representation of trust to per-ceive the state of the cooperation context and to allowthe user to better adapt his action to the context. Theanalysis of trust is made during the activity. In the caseof the use of a risk management approach, the objec-tives are identifying the types of risk, evaluating them,establishing a scale of risks, defining scenarios to reduceand correct the risks, and finally monitoring risks dur-ing the activity (Alquier and Tignol, 2007). This secondmethod requires a consistent study, whereas the methodpresented here suggests dynamically using the informa-tion of the cooperation context to provide automaticallytrust indicators in the correct progression of the activity.

The trust criteria were identified during stages ofbrainstorming with researchers. Then 14 professionals


Table 1Trust criteria and information sources

(1) Task Progress-Specific Trust Information sources

State of the task Gantt Planning, PertPlanning, 4D

Problems of execution Remarks in the meetingreport

Environment Weather forecast, etc.

(2) Actor-Specific Trust Information sources

Competence Certification (e.g., ISO)Performance Performance evaluation

systemAttendance at building site

meetingMeeting report

(3) Building Element-SpecificTrust

Information sources

Difficulty level of execution Technical report, standardnorms

Modifications List of modificationsRespect of budget Budget monitoring

(4) Document-Specific Trust Information sources

State of the documents List of documentsState of the documents

requestsDocuments transmission list

Availability on the buildingsite

List of documents

from the construction sector (architects, engineers, andcontractors) were confronted with the results to ad-just and validate these findings. Table 1 considers fourcategories of Specific Trust associated with the four as-pects of the activity (i.e., task progress, actor, buildingelement, and document) and presents the final results(criteria and sources).

1. The assessment of “Task Progress-Specific Trust”(Ttp) depends on:• The state of the task. Trust relies on similarity

with a state initially planned or with a state thatcould be more favorable (e.g., building construc-tion activity in advance).

• Problems identified during the execution. A highlevel of remarks coming from the building sitemeeting report and corresponding to the taskunder consideration leads to decrease the trustlevel.

• The environment. Trust is determined by the riskrun by the task when there is bad weather fore-cast, or a potential seism.

2. The assessment of “Actor-Specific Trust” (Ta) iscoming from:• The competence. It is materialized through var-

ious certifications (e.g., ISO). The fact thatan organization is certified constitutes a goodwarrantee.

• The performance. It is coming from previ-ous experience and it is possibly capitalizedin some performance evaluation systems (e.g.,Rating Source see http://www.ratingsource.comfor more information). Positive previous expe-rience contributes to increase trust, while nega-tive previous experience contributes to increasemistrust.

• Attendance at building site meeting. Stakehold-ers have to regularly attend construction sitemeetings to be informed about the state of theconstruction activity and about all the decisionsconcerning the activity that they have to per-form. If stakeholders do not frequently attendthe building site meetings, they cannot be in-formed about decisions and, therefore, this canbe a source of dysfunction.

3. The assessment of “Building Element-SpecificTrust” (Tbe) relies on:• The difficulty level of the task execution. It

comes from the construction manager’s experi-ence or from other technical judgment issuedby organisms such as the CSTB (Scientific andTechnical Centre for the Building Industry,French organism) for more complex or experi-mental technical design. A weak level of diffi-culty of task execution ensures a high trust level.

• Modifications in comparison with the specifica-tions. The modifications of the building elementduring the construction activity are a sourcefor disputes between the owner and contrac-tor concerning the issue of compensation forchanges (Karim and Adeli, 1999a) and also asource of mistakes at the moment of the execu-tion. Indeed, a modification on a building ele-ment implies taking into account its impact onthe whole building project and on all the docu-ments. Building elements on which we can iden-tify some modifications have to be more con-trolled and consequently considered as having aweak trust level.

• Respect of budget. Modifications on building ele-ments and unexpected events (e.g., delay linkedto the nature of the ground) are sources of diffi-culty to respect the budget. These types of situ-ations have to be more controlled.

4. The assessment of “Document-Specific Trust” (Td)required for the task takes into account:


Fig. 4. Approach for measuring Global Trust and Specific Trust.

• The state of the documents. The fact that allthe documents are up-to-date contributes to in-creasing the trust level.

• The state of the documents requests. The re-quests linked to the documents (e.g., request formodification, for validation, etc.) have to be per-formed within the time limits allowed. A delaycontributes to decreasing the trust level.

• The availability on the building site. The docu-ments required for the execution of the buildingelement have to be available on the building site.The trust level increases if all the required doc-uments are available on the building site.

4.2 Measure of the trust indicators

To evaluate the trust level in the correct progressionof an activity, we have adapted S.P. Marsh’s approach(Marsh, 1994), which identifies the trust during cooper-ation to our specific cooperation context.

Two levels of trust indicators can be distinguished:the Global Trust and the Specific Trust. These indica-tors can be automatically measured on the basis of theknowledge of the cooperation context centralized in aninformation system that allows identifying the value ofthe criteria describing before.

Global Trust characterizes “trust in the correct pro-gression of the activity” in a particular situation. SpecificTrust corresponds to each aspect of the task. We distin-guish between four types of Specific Trust in a particularsituation (as seen in Section 4.1):

• Task Progress-Specific Trust,• Actor-Specific Trust,• Building Element-Specific Trust,• Document-Specific Trust.

Figure 4 illustrates our approach for measuringGlobal and Specific Trust.

Table 2Measure of Global Trust: summary of annotations

Description Representation Value range

Activity α

Global Trust T(α) [−1, 1]Number of Specific Nx [0,n]

Trust categories x = tp, a, d, beType x Specific Trust Txi (α) [−1, 1]

in the situation α

Importance of the type Ixi (α) [0,1]x in the situation α

4.2.1 Measure of Global Trust. T(α) represents GlobalTrust within the correct progression of an activity α

(Table 2 summarizes the notions).

T(α) =

Nx∑

i=1

Txi (α)Ixi (α)

Nx∑

i=1

Ixi (α)

Thus, Global Trust results from trust in each as-pect of the task (progress of the task under considera-tion, actors in charge of its execution, building elementsresulting from the task, required documents). Theweight (Ixi (α)) denotes the importance of each indica-tor with respect to Global Trust. This weight is valuatedby the building construction manager. Marsh (1994) hasestablished that “importance is an agent-centered or sub-jective judgment of a situation on the part of the agentconcerned.”

4.2.2 Measure of the Specific Trust. The values ofTtp(α), Ta(α), Tbe(α), and Td(α) have to be determined.These values depend on the intrinsic characteristics of


Table 3Measure of Specific Trust: summary of annotations

Description Representation Value

Set of criteria for type x Specific Trust Cx Set of AlphanumericType of Specific Trust x tp; a; be; dCriterion i for type x Specific Trust Cxi 0;1Number of criteria for type x Specific-Trust nx [0,n]Set of values of Criterion i for type x Specific Trust V(Cxi ) Set of AlphanumericSet of positive values of Criterion i for type x Specific Trust V+(Cxi ) Set of AlphanumericSet of negative values of Criterion i for type x Specific Trust V−(Cxi ) Set of AlphanumericSet of null values of Criterion i for type x Specific Trust V0(Cxi ) Set of AlphanumericNumerical of i for type x Specific Trust in situation α NV(Cxi )α −1; 0; 1Value of Criterion i for type x Specific Trust in situation α V(Cxi )α AlphanumericType x Specific Trust in situation α Tx(α) [−1,1]Coefficient of Criterion i of type x Specific Trust γ (Cxi ) [0,1]

the different aspects of the task. To evaluate the fourtypes of trust, the following formula is used.

Tx(α) =nx∑

i=1

NV(Cxi )α ∗ γ (Cxi )

Cx = {Cxi } x ∈ {tp, a, be, d} et i ∈ [1, nx]

NV(Cxi )α = 1 si V(Cxi )α ∈ V+(Cxi )

NV(Cxi )α = −1 si V(Cxi )α ∈ V−(Cxi )

NV(Cxi )α = 0 si V(Cxi )α ∈ V0(Cxi )

γ (Cxi ) = 1nx

Table 3 summarizes the annotations.This method suggests that each trust criterion has

the same weight in the calculation. Even if the inser-tion of a weight factor could refine the resulting values,it seems more important to limit the individual settingto limit the information coming from the constructionmanager.

4.3 Application of the method to measure trustindicators to a scenario

For example, we consider the particular activity“Platform-First Part.” We have to identify data relatedto the task and the criteria necessary for the measureof trust (see Table 4). Then, we associate the criterionvalue to a numerical value:

• [+1]: if the criterion value is a positive value for thecorrect progression of the task.

• [−1]: if the criterion value is a negative value for thecorrect progression of the task.

• [0]: if the criterion value is a neutral value for the cor-rect progression of the task (e.g., if we do not havedata related to the criterion).

So, if we consider the task “Platform-First Part,” wecan identify the following situation:

• The progress of the task conforms to the planning.• There are some remarks related to the task in the

building site meeting report.• The weather forecast is unfavorable for performing

the task.• The contractor has a competence certification (Qual-

ibat, French certification for contractors).• Previous project experiences were positive.• The contractor frequently attends the building site

meetings.• The building element is not particularly difficult to

construct.• There is no modification in comparison with the spec-

ifications for the building element.• The budget is respected for this building element.• The documents are good for execution, there are no

more open requests but one document is not availableon the building site.

Table 4 summarizes the situation and explains themeasures of Global and Specific Trust.

5 TRUST-BASED DASHBOARD FOR THECONSTRUCTION MANAGEMENT

5.1 Functional description of the prototype

As seen in Section 1, the construction manager usesmultiple views to coordinate the building construc-tion activity (e.g., planning, meeting report, etc.). For-mer research works allowed establishing relationshipsbetween the contents of these views. These workscontributed to implement an infrastructure of models


Table 4Measure of Global and Specific Trusts: example task “Platform-First part”

Measure of Specific Trust

(1) Task Progress-Specific Trust – Ttp(α) (2) Actor-Specific Trust –Ta(α)

Criterion V NV Criterion V NV

State of the task In progress +1 Competence certification Yes +1Problems of execution 3 remarks −1 Performance Positive +1Environment −10 ◦C −1 Attendance at building site meetings Frequent +1

Ttp(α) = −0.33 Ta(α) = 1

(3) Building Element-Specific Trust – Tbe(α) (4) Document-Specific Trust – Td(α)

Criterion V NV Criterion V NV

Level of difficulty of execution Weak +1 State of the documents Good for execution +1Modifications No +1 State of the documents requests Finished +1Respect of budget Ok +1 Availability on the building site No −1

Tbe(α) = +1 Td(α) = +0.33

Measure of Global Trust

Global Trust – T(α)

Specific Trust NV Importance

Task Progress-Specific Trust −0.33 +1Actor-Specific Trust +1 +0.75 T(α) = +0.47Building Element-Specific Trust +1 +1Document-Specific Trust +0.33 +1

based on the concepts of Model-Driven Engineering(MDE) (Favre, 2004; Bezivin, 2005) to describe the co-operation context model in the AEC sector (Kubickiet al., 2007b). They also led to the implementationof a prototype called “Bat’iViews,” further detailed in(Kubicki et al., 2006). This prototype suggests puttinginto relation the concepts of four dynamic views insidea “multiviews interface” (Wang-Baldonado et al., 2000).These views, which are regularly manipulated by thestakeholders of the building construction activity, arethe following ones: (1) meeting report, (2) 3D model,(3) planning, and (4) list of remarks in the meeting re-port. The user can freely navigate inside these views.When he selects an element within a view (e.g., a taskin the planning), Bat’iViews highlights the correspond-ing information in the other views (e.g., the building ele-ments corresponding to the task under consideration inthe 3D model).

Our proposal of a dashboard based on trust is inte-grated in the continuity of these works. The Bat’iTrustprototype suggests inserting a new view: a dashboardbased on trust intended for the construction manager.This dashboard allows visualizing the trust indicatorsautomatically calculated on the basis of the method de-tailed in Section 4.2. Its objective is to improve the per-ception of the state of the cooperation context. It alsoaims to identify the potential dysfunctions on the build-

Fig. 5. View of the trust-based dashboard.

ing site. The interface of the dashboard (see Figure 5)displays a list of construction tasks, their state (e.g., onhold, in progress, etc.), and the following types of trustindicators:


Fig. 6. Principle of use of the Bat’iTrust prototype.

• Global Trust Indicator,• Task Progress-Specific Trust Indicator (TP-STI),• Actor-Specific Trust Indicator (A-STI) (it concerns

actors in charge of the execution of the task underconsideration),

• Building Element-Specific Trust Indicator (BE-STI)(it concerns building elements resulting from thetask),

• Document-Specific Trust Indicator (D-STI) (it con-cerns documents required for the execution of thetask).

This trust-based dashboard provides an entry pointin the navigation for the Bat’iTrust users. The selectionof a specific trust indicator returns a particular config-uration of views that enables the understanding of thesituation. It also highlights information about the con-struction task under consideration. For example, the se-lection of a progress-specific trust indicator displays aconfiguration of views composed of a planning, a viewof the remarks (coming from the meeting report), and aview of weather forecast. Figure 6 identifies the config-

uration of views suggested for each type of specific trustindicators.

Therefore, the use of Bat’iTrust allows the construc-tion manager to rapidly identify potential dysfunctionsand easily understand the nature of dysfunctions. Af-ter that, he can take decisions and adapt his action tothe cooperation context. For example, we consider thescenario detailed in Figure 7. The construction manageridentifies a weak level of trust in the task “Platform-First part” for the “task-progress” dimension. He selectsthe task progress-specific trust indicator. Bat’iTrust re-turns a specific configuration of views and highlights ineach view the pieces of information corresponding tothe task under consideration. The construction man-ager sees in the planning that the task is late. Thereis one remark about this task in the meeting reportbut the weather forecast is favorable for the task underprogress. The prototype allows the construction man-ager to have at his disposal information to understandthe nature of the dysfunction related to task-progress,and consequently, it allows him to better identify whatkind of action he has to perform to limit the impact of


Fig. 7. Scenario of use of the Bat’iTrust prototype.

the dysfunctions (e.g., adjust the planning, or suggest toincrease the available human resources).

5.2 Implementation

The Bat’iTrust prototype has been implemented on thebasis of functional specifications and views modeling.When designing a dashboard helping to monitor a col-lective activity (i.e., a construction project), one majorquestion remains in the sources of information to beconsolidated. Trust criteria refer to several sources ofinformation as described in Table 1. But informationabout the activity is largely distributed between multi-ple sources (e.g., documents and information systems).This point is important because it often leads to a lackof coherence between information sources but also to alack of shared semantics and interoperability betweenIT systems.

5.2.1 Cooperation context and service-based architec-ture. To surmount these limits, we have based our ap-

proach on a standard cooperation model. The imple-mentation of the Bat’iTrust prototype relies on a sharedmodeling of the cooperation context enabling to man-age a reliable source of information about the ongoingcollective activities (i.e., the construction projects). Spe-cific cooperation context models, taking into account thecharacteristics of specific AEC projects, are instantiatedfrom a cooperative activity meta-model.

This approach is further discussed in (Kubicki et al.,2006). It allows conceptually representing informa-tion related to a construction project. Therefore, ITapplications used by the practitioners are considered asinputs of data feeding the cooperation context informa-tion system.

The design of applications supporting cooperationprojects is addressed in several RDI projects. InLuxembourg, the CRP Henri Tudor has developedthree IT applications and their related business services(Kubicki et al., 2007a, 2009):

• A shared application for the management of meet-ing reports allows the coordinators to write and share


Fig. 8. Services-based cooperation context management.

their remarks about the construction process aftera team meeting. Each participant of the projectcan then access the remarks through a web-basedapplication. This application provides the coopera-tion context with information related to the dysfunc-tions on building site and to the progress of thetasks.

• A shared application for document management en-ables the sharing of documents in the constructionproject. It implements several services related to thedocument management such as uploading, download-ing, adding request and comments about documents,and so on. This application provides information re-lated to the exchanges of documents produced andtheir validations.

• A task management application allows the coordina-tor to manage the planning by importing standardplanning files (e.g., MS Project). Each week he canprecisely trace the progress of each task.

These applications are designed as “input of infor-mation” in the shared model of the cooperation con-text described previously (see Figure 8). Then, beingused on several “real life” projects, they allow ben-efiting of representative data sets about constructionprojects.

They are designed according to service-based tech-nologies permitting to clearly distinguish between thecooperation context itself (i.e., a shared and coordi-nated source of information about the activity) andthe IT applications using it. This approach allows tak-

ing into account the numerous existing business ap-plications, used by one or more actors of the project.Very often these applications only manage a part (aview) of the cooperation context. For instance a doc-ument sharing application will only manage a part ofthe cooperation context related to document exchangesbetween participants of a project (e.g., document up-load/download, versioning, requests between users). Inour service-based approach, these specific applicationscan make use of business services to be connected tothe shared cooperation context, and therefore inter-act with project information (i.e., store and retrievedata).

The Web services are described in the REST protocol(Fielding, 2000) but should also be available in SOAP,which is most common in the IT developers’ commu-nity. REST is a Web services technology based on theWeb architecture and its basic technologies: HTTP,URI, and XML. We have structured these Web servicesusing the Resource Oriented Architecture (ROA) ap-proach (Richardson and Ruby, 2007). ROA describesa set of good practices for REST Web service designand is very adapted to our Agile development process(Abrahamsson et al., 2002), involving business experts,technical experts, and final users in the design process.

5.2.2 Bat’iTrust prototype. The Bat’iTrust prototype(see Section 5.1) requires a coherent and valid set ofproject data to automatically calculate the global andspecific trust indicators and to display the content of the


views. Therefore, the proposal relies on the cooperationcontext information system in combination with third-parties service-based applications feeding it.

The Bat’iTrust application is composed of a graphicalmultiviews interface. As described in Section 5.1, a mas-ter view (the dashboard) allows the user to interact withthe other views. We have implemented this multiviewsinterface using Rich Internet Application (RIA) tech-nologies, making it available for the user with a standardWeb browser. We have chosen the Adobe Flex solution,providing fast results useful in our prospective approach(validation of the concepts with practitioners). Flex alsoprovides an integrated development environment gen-erating Flash code directly.

The application has been developed using a Model,View, Controller (MVC) architecture enabling one toclearly separate the data (i.e., model), the visualiza-tion, and the controls. This approach is closely in-tegrated within the service-based cooperation contextmanagement described before. The cooperation con-text provides data by using Web services. Views of theBat’iTrust interface make use of these services in func-tion of their visualization model (see Section 5.2). Theinteractions between views (i.e., arrangement of viewswhen a specific trust indicator is selected) are also man-aged by Web services providing the conceptual relation-ships between the displayed business concepts.

6 VALIDATION OF THE PROPOSITION

This experiment has been made in collaboration withresearchers and professionals of the construction sector.Until now, seven people have tested the Bat’iTrust pro-totype. The objectives have been the following:

• Validation of the potential of the trust representationto assist the construction management.

• Validation of the contents of the different configu-rations of views to analyze the dysfunctions on thebuilding site.

6.1 Experiment protocol

In the course of the experiment, data have been largelyinspired by a real construction project to obtain a con-text close to an actual context of building construc-tion activity. The experimental scenario has been simu-lated but it relies on information collected during a realschool construction project. The experiment protocol isdivided into three sequences:

• During the first stage, the experimenter has to con-sult and manipulate paper documents usually used tomanage the construction activity (i.e., meeting report,

pictures, list of documents). A problematic scenariois highlighted in these documents to guide him inthe understanding of the dysfunction, based on theseusual documents.

• Then, the second stage is intended for the use of theBat’iTrust prototype. The user has to navigate in thedashboard and in the different views to understandthe dysfunction. We make use of software to capturethe manipulations on the screen and to record the ex-perimenter’s reactions (Web cam, audio).

• Finally, the experimenter has to fill in a survey ques-tionnaire about the utility and the usability of theprototype.

6.2 User feedback

The experiment stage allowed us to collect the firstusers’ feedback. Seven people (i.e., three architects, twoarchitect researchers, and two students) served as ex-perimental subjects. The results of the questionnaire(see Figure 9 for an extract of user feedback analysis)allowed assessing the relevance of this proposal. Usersfind the dashboard based on trust efficient to guidetheir navigation and the trust indicators interesting forthe construction management (even if some of themprefer staying neutral because they have not used themin a real project context). Moreover, they think that itallows easily, putting into relation the content of the di-verse views. They also believe that it is an interestingtool to coordinate and monitor the construction activityand each of them would like to use the tool in their ownconstruction projects.

Some users highlight that the device displays a largeamount of information. Nevertheless, this characteristicdoes not limit the navigation because information visu-alization modes are similar to those used in the profes-sional practices (e.g., planning, weather forecast, etc.),and consequently, do not require a long training.

However, discussions with users emphasize the factthat they regret not being able to see the detail of trustcalculation for the global and specific trust indicators.The resulting value is interesting, but without the de-tailed calculation some users have difficulty in interpret-ing the indicators. Therefore, we are now working on anew functionality in our dashboard to access this type ofinformation.

This experiment stage has allowed demonstrating theinterest of this proposition for the construction sec-tor. The indicators of trust allow efficiently structuringthe navigation in the coordination assistance tool andhighlighting the potential dysfunctions on the buildingsites. Then user feedback allows deducing that the con-tents of the configurations of views are adapted for the


Fig. 9. Users feedback: extract of results.

business practices. They allow understanding the natureof the dysfunctions and supporting the decision pro-cess. Consequently, trust representation (i.e., trust in-dicators) is a new interesting approach of the visual-ization of the cooperation context. The combination ofthese trust indicators with the specific configurations ofviews constitutes a suitable support for the construc-tion management. That is why the users have expresseda favorable opinion to include these prototype andmethod of construction management in their everydaypractice.

The first results obtained during this experiment stageare encouraging but they are not sufficient to com-pletely validate the proposal mainly because of the lim-ited number of experimental subjects. Moreover, theexperiment period is short (i.e., +/− 40 minutes) andthe experimental scenario is a simulated context rely-ing on a real building construction activity. Neverthe-less, this stage of experiment consists in a first step ofvalidation and allows considering new implementationperspectives (e.g., introduction of a detailed view of thetrust indicators calculation).

7 PROSPECTS AND CONCLUSION

In conclusion, the mission of the construction man-ager has been presented and the potential of a dash-board view based on trust as a coordination assistancetool has been highlighted. At this stage of this researchwork, some construction managers have been inter-

viewed, and a survey has been diffused to the profes-sionals of the construction sector (architects, engineers,and contractors) to validate and to adjust our trust cri-teria. A method for the calculation of trust in the cor-rect progression of the activity has been proposed. Afirst Bat’iTrust prototype has implemented this methodand has been exposed to professionals from the AECsector, to assess the relevance of the proposal and toadjust it according to feedback. It is developed on aservice-based infrastructure allowing for the manage-ment of AEC projects’ cooperation context. The aim isto make the prototype as generic as possible (i.e., in-dependent from a given suite of IT tools). Service ap-proach seems to be a way to achieve this. Bat’iTrust isnow an independent visualization Flex application usingREST Web services to access cooperation context data.Technically, it should be connected to other sourcesof information (e.g., Primavera IT solutions for AECprojects) through the reuse (or design) of application-specific services.

Finally, a first stage of experiment has been con-ducted to assess the relevance of trust to support con-struction management. Even if there are some limitsconcerning this stage of experiment (e.g., limited num-bers of experimental subjects, limited period of use ofthe device, experiment based on a simulated scenario),trust appears as having potential in this context. Theusers feedback is encouraging and lets us demonstratethat trust and its representation can support the coor-dination of building construction activity and can be atthe origin of a new generation of Web-based construc-tion management tools.


ACKNOWLEDGMENT

The authors would like to thank the MCESR (Research,Higher Education, and Culture Ministry) and the FNR(National Research Fund) in Luxembourg, which havefunded this research work.

REFERENCES

Abrahamsson, P., Salo, O., Ronkainen, J. & Warsta, J. (2002),Agile software development methods. Review and analysis,University of Oulu, Oulu, Finland.

Adeli, H. & Karim, A. (1997), Scheduling/cost optimizationand neural dynamics model for construction, Journal ofConstruction Engineering and Management, ASCE, 123(4),450–8.

Alquier, A.-M. B. & Tignol, M.-H. L. (2007), Managementde Risques et Intelligence Economique, L’approche Prima,Ed. L’intelligence economique, Paris, Economica.

Arslan, G., Kivrak, S., Birgonul, M. T. & Dikmen, I. (2008),Improving sub-contractor selection process in construc-tion projects: Web-based sub-contractor evaluation system(WEBSES), Automation in Construction, 17(4), 480–88.

Bezivin, J. (2005), On the unification power of models, Soft-ware and Systems Modelling (SoSym), 4(2), 171–88.

Bobroff, J. (1994), La gestion de projet dans la construction,Ed. Presses de l’Ecole Nationale des Ponts et Chaussees,Enjeux, organisations, methodes et metiers, Paris.

Boone, A. (2007), Gestion des risques dans la construction,Les dossiers du CSTC. N◦ 2/2007, Cahier n◦ 1. CSTC/BBRI,Bruxelles.

Chang, E., Dillon, T. & Hussain, F. K. (2006), Trust and Repu-tation for Service-Oriented Environments, Technologies forBuilding Business Intelligence and Consumer Confidence(ed.), John Wiley & Sons Ltd., Chichester, UK.

Chau, K., Anson, M. & Zhang, J. (2005), 4D dynamic con-struction management and visualization software, Automa-tion in Construction, 14(4), 512–24.

Chemillier, P. (2003), Demarche qualite dans les entreprisesdu batiment, Ed. Saint-Denis La Plaine, AFNOR.

Chi, S., Caldas, C. H. & Kim, D. Y. (2009), A method-ology for object identification and tracking in construc-tion based on spatial modeling and image matchingtechniques, Computer-Aided Civil and Infrastructure Engi-neering, 24(3), 199–211.

Deutsch, M. (1962), Cooperation and trust: some theoreticalnotes, in M. R. Jones (ed.), Nebraska Symposium on Mo-tivation, University of Nebraska Press, Nebraska, pp. 275–320.

Ekstrom, M. (2004), Accounting for Rater Credibility whenEvaluating Construction Industry Service Providers, CIFETechnical Report # 148, Stanford University, Stanford, CA.

Favre, J. M. (2004), Towards a Basic Theory to Model DrivenEngineering, 3rd Workshop in Software Model Engineer-ing, WISME 2004.

Fielding, R. T. (2000), Architectural Styles and the Design ofNetwork-based Software Architectures, Ph.D. thesis in In-formation and Computer Science, University of California,Irvine, CA, 162 pp.

Gachet, A. & Brezillon, P. (2005), A context-based represen-tation of knowledge flows in dynamic organizations, apply-

ing context management, RSTI Revue d’Intelligence Artifi-cielle, 19(3), 557–74.

Halin, G. & Kubicki, S. (2008), Une approche par les modelespour le suivi de l’activite cooperative de construction d’unbatiment, Une interface multi-vues et des services metiersorientes gestion de chantier, Revue ISI (Ingenierie desSystemes d’Information), Modeles, Formalismes et Outilspour les Systemes d’Information, 13(4), 35–58.

Hardin, R. (2000), Trust and Trustworthiness, Ed. RussellSage Foundation, New York.

Jin, X. L. & Doloi, H. (2009), Modelling risk allocationin privately financed infrastructure projects using fuzzylogic, Computer-Aided Civil and Infrastructure Engineer-ing, 24(7), 509–24.

Karim, A. & Adeli, H. (1999a), Object-oriented informa-tion model for construction project management, Jour-nal of Construction Engineering and Management, 125(5),361–7.

Karim, A. & Adeli, H. (1999b), CONSCOM: An OO con-struction scheduling and change management system, Jour-nal of Construction Engineering and Management, 125(5),368–76.

Klemetti, A. (2006), Risk management in construction projectnetworks, Laboratory of Industrial Management, HelsinkiUniversity of Technology, Helsinki, Finland.

Kramer, R. M. (1999), Trust and distrust in organizations:emerging perspectives, enduring questions, Annual Reviewof Psychology, 50, 569–98.

Kubicki, S., Bignon, J. C., Halin, G. & Humbert, P. (2006),Assistance to building construction coordination. To-wards a multi-view cooperative platform, ITcon ElectronicJournal of Information Technology in Construction,11(Special Issue “Process Modelling, Process Managementand Collaboration” edited by P. Katranuschkov), 565–86.

Kubicki, S., Dubois, E., Halin, G. & Guerriero, A. (2009),Towards a sustainable services innovation in the construc-tion sector, in Proceedings of CAISE, 21st InternationalConference on Advanced Information Systems, LNCS 5565,pp. 319–33.

Kubicki, S., Guerriero, A., Hanser, D. & Halin, G. (2007a),IT services design to support coordination practices in theLuxembourguish AEC sector, in Proceedings of the 4th Co-operative Design Visualization and Engineering Conference,396–403.

Kubicki, S., Halin, G. & Guerriero, A. (2007b), Multi-visualization of the cooperative context in building con-struction activity. A model-based approach to designAEC-specific visualization interfaces, in Proceedings ofBuiltViz’07, International Conference on Visualisation inBuilt and Rural Environments, 590–5.

Luhmann, N. (1988), Familiarity, confidence, trust: problemsand alternatives, in Diego Gambetta (ed.), Trust: Makingand Breaking Cooperative Relations, Basil Blackwell, NewYork, pp. 94–107.

Marsh, S. P. (1994), Formalising trust as a computational con-cept, Ph.D. dissertation, University of Stirling, Departmentof Computing Science and Mathematics, Stirling, Scotland,170 pp.

Mayer, R. C., Davis, J. H. & Schoorman, F. D. (1995), An in-tegrative model of organizational trust, Academy of Man-agement Review, 20(3), 709–34.

Rebolj, D., Babic, N. C., Magdic, A., Podbreznik, P. &Psunder, M. (2008), Automated construction activity


monitoring system, Advanced Engineering Informatics,22(4), 493–503.

Rekapalli, P. V., Martınez, J. C. & Kamat, V. R. (2009), Algo-rithms for accurate three-dimensional scene graph updatesin high speed animations of simulated construction oper-ations, Computer-Aided Civil and Infrastructure Engineer-ing, 24(3), 186–98.

Richardson, L. & Ruby, S. (2007), RESTful Web Services,O’Reilly, Sebastopol, CA.

Rousseau, D. M., Sitkin, S. B., Burt, R. S. & Camerer, C.(1998), Not so different after all: A cross-discipline view oftrust, Academy of Management Review, 23(3), 393–404.

Sadeghpour, F., Moselhi, O. & Alkass, S. (2004), A CAD-based model for site planning, Automation in Construction,13, 701–15.

Senouci, A. B. & Adeli, H. (2001), Resource scheduling us-ing neural dynamics model of Adeli and Park, Journalof Construction Engineering and Management, 127(1), 28–34.

Simon, H. A. (1959), Theories of decision-making in eco-nomics and behavioral science, American Economic Re-view, 49, 253–83.

Suchman, L. A. (1987), Plans and situated action: The prob-lem of human-machine communication, Ed. R. Pea et J. S.Brown, Cambridge University Press, Cambridge, UK.

Sutcliffe, A. (2006), Trust : from cognition to conceptual mod-els and design, in Proceedings of the 18th Conference on Ad-vanced Information Systems Engineering, 3–17.

Tahon, C. (1997), Le pilotage simultane d’un projet de con-struction, Ed. Plan Construction et Architecture, Col-lection Recherche n◦87, Paris, Plan Construction etArchitecture.

Uden, L. & Naaranoja, M. (2007), The development of on-line trust among construction teams in Finland, ITcon Elec-tronic Journal of Information Technology in Construction,12, 305–21.

Wang-Baldonado, M. Q., Woodruff, A. & Kuchinsky, A.(2000), Guidelines for using multiple views in informationvisualization, in Proceedings of the Working Conference onAdvanced Visual Interfaces, 110–19.

Zolin, R., Levitt, R. E., Fruchter, R. & Hinds, P. J. (2000),Modeling & Monitoring Trust in Virtual A/E/C Teams, AResearch Proposal, Stanford University, Stanford, CA.

Zou, P. X. W., Zhang, G. & Wang, J.-Y. (2007), Under-standing the key risks in construction projects in China,International Journal of Project Management, 25(6), 601–14.

Zucker, L. (1986), Production of trust: institutional sources ofeconomic structure: 1840–1920, Research in OrganizationBehaviour, 8, 53–111.