evaluating the cumulative impact of changes on labour productivity

C hanges on a construction projectare commonplace, despite thefact that contractors often signfixed price contracts. This may be due inpart to the fact that the owners often have aunilateral right in the contract to imposesuch change. Reasons for changes mayinclude new requirements, design errorsand omissions, inadequate contractdocuments, and differing site conditions, toname a few.

Almost every change to a constructionproject has some effect on the projects costand schedule. There are generally twocategories of this effect:

The actual direct cost and time ofperforming the change (additionallabor, materials, equipment). And,

The impact the change may have onother unchanged or contractual workbecause of delay, disruption, change ofsequence, lack of resources, etc.

Contractors generally do not havemuch difficulty estimating the direct costsor time required to perform a single change,but it is very difficult to accurately estimatethe impact of that change on theunchanged or contractual part of the work.To execute a change it may be necessary todisrupt or delay the unchanged work, orperform it in a manner or sequencedifferent to that originally planned, all ofwhich may lead to a loss of productivity andincreased costs.

Difficulty in estimating the impacts,coupled with the resistance on the part of

the owner to recognize such impacts, oftenleads to a decision to leave impacts out of achange order. Neglecting impacts may notbe a big problem if the number and value ofthe changes on a project is minimal, but thesituation becomes much more complex inthe case of hundreds if not thousands ofchanges on a large project.

When numerous changes occur, thereis a compounding and negative effectcommonly called cumulative impact.This impact is poorly understood, difficultto measure, and seldom reflected in thesum of the estimated costs of individualproject changes. The additional or changedwork is often executed by means such asshort-term hiring, overtime and doubleshifts. There can be disruptions and delays,loss of learning, trade stacking, materialsourcing problems, low morale, all of whichlead to significantly lower levels ofproductivity.

This compounding and negative effectoften goes unnoticed until it is too late. Itgenerally becomes apparent only in thelatter stages of a project when work cannotbe completed on time and when laborproductivity does not measure up to theanticipated levels.

The difference between actual hoursand estimated hours cannot, however, beexplained simply by the number ofapproved change order hours. The balancerepresents unexplained extra hours whichmay partly be the result of anunderestimate, the contractors owninefficiencies, or, in the case of a project

suffering a multitude of changes, the lossmay be the impact of the change orders.

Most project mangers, when looking atan individual change, also tend to beoptimistic about their ability to incorporatethe change without otherwise disruptingthe project. When pricing numerouschanges, the contractor fails to foresee, andthe owner fails to acknowledge, that theimpact on the project as a whole may begreater than the sum of the hours workedon the individual changes. Furthermore,because of pressure from owners and theiragents, project managers will allow latechanges to be introduced, even on projectsthat are trending behind schedule or overbudget, further aggravating the situation.The later a change occurs on a project, theless efficiently it is implemented.

Determining the impacts that changescan have on contract price and time can bearduous. As a result, it is difficult for ownersand contractors to agree on equitableadjustments, especially for cumulativeimpact. Consequently, projects that havesuffered a multitude of changes and exceedcost or schedule targets are likely to lead tocumulative impact disputes. A simplemethod for estimating the possible impactwould be invaluable.

EVALUATING THE IMPACT

Owners frequently think thatcontractors make money on changesbecause their estimates are too high. Inreality, contractors often lose money onmultiple changes because their estimatesare too low and because they underestimatethe administrative effort required tonegotiate and process the change. Pricingmethodologies for changed work are oftenweak. Few contractors maintain adequatejob-site records to allow evaluation ofimpact costs for individual change orders,let alone a multitude of change orders.

With each individual change, acontractor will estimate the work-hoursrequired, but because of the inability ofproject personnel to fully anticipate theconsequential effects of multiple changes,the actual final work-hours required may bemuch greater than originally anticipated. Asthe number of changes increases, thedifferential between estimated work-hoursand actual work hours widens at anincreasing rate.

Over the last 20 years, a number ofstudies have been published that attempt to

Cost Engineering Vol. 50/No. 12 DECEMBER 2008 23

TECHNICAL ARTICLE PEER REVIEWED

Evaluating the Cumulative Impactof Changes on Labor Productivityan Evolving DiscussionDr. William Ibbs and Gerald McEniry, CFCC PSP

ABSTRACT: This article presents the results of a meta-analysis of two studies and shows in quan-titative terms that increasing amounts of project change will have significant impact on laborproductivity. The result is a series of curves that can be used as guidelines for both estimatinglabor productivity before and after change event(s). Two standards have emerged from work byCharles A. Leonard in 1988, and by William Ibbs in 1995, 1997, and 2005. Compiled fromresearch studies of actual construction projects, these studies developed a statistical relation-ships between the amount of change and end-of-project labor productivity. Though not per-fect, they are more impartial, clearly defined, and easier to use.

KEY WORDS: Change orders, construction, labor productivity, overtime, projects, and standards

evaluate this impact. Two standards haveemerged from early work by Charles A.Leonard in 1988 [11] and more recently byWilliam Ibbs in 1995, 1997, and 2005 [6, 7,8]. Compiled from research studies ofactual construction projects these studiesdeveloped statistical relationships betweenthe amount of change and end-of-projectlabor productivity. This article will focusprimarily on these two studies, yet will alsoprovide some comments on other methodsas they relate to these studies.

Leonard MethodLeonards work was the first published

study to test for a statistical correlationbetween change orders and productivity[11]. Using data gathered from 90construction disputes arising on 57 differentprojects while working for a claimsconsulting firm, he parsed the informationinto two different graphs, one forcivil/architectural contracts and the otherfor electrical/mechanical work.

Figure 1 is representative, and showsthree curves for his electrical/mechanicalprojects. The lowest curve is a linearregression for those projects in his databasethat were substantially affected by changeorders only. The other two curves representprojects that were substantially impacted bychange orders and one or more majorcauses of productivity loss such asinadequate scheduling and coordination,acceleration, change in work sequence, latesupply of information, equipment ormaterials and increased complexity of work.His other diagram, which is not shown herefor reasons of space, is for architectural/civilprojects and also has three parallel curves.

As illustrated in figure 1, Leonardchose to illustrate the impact of the changesin terms of percent lost productivity (LOP)which is the ratio of the unproductive laborhours to the actual labor hours spent on theoriginal contract (i.e., unchanged work).Most researchers have chosen to illustratethe impact of changes in terms of theconstruction productivity index (PI)defined in equation 1:

LOP = (1-PI) x 100%, where PI = earnedhours actual hours

(equation 1)

Figure 1 can be transformed into figure2 for the purpose of comparing Leonardsresults with other research.

Although different in appearance, bothfigures 1 and 2 are basically the same sincethe loss of productivity is related to theproductivity index as mentioned earlier. Italso should be noted that figure 2 includescertain outlying data points, at values of lessthan 10 perent change, discarded byLeonard from his final graph. He cautionedthat his results were only good for change inthe 10 percent to 60 percent range.

The percent change order factor iscalculated by dividing the change orderlabor-hours by the actual contract labor-hours. Actual contract labor-hours aredefined by Leonard to be the total actualhours on the project less change orderhours less any hours attributable to thecontractors own inefficiencies or bidmistakes, see equation 2.

Percent change orders = change orderlabor-hours / (total actual labor-hours -change order hours - contractor mistakes)

(equation 2)

As an example calculation, consider acase where the normal or earned number ofhours required for a project was determinedto be 10,000 hours but that 20,000 hourswere actually spent, including 6,000 hoursworked and paid on change orders.Assuming for simplicity that no inefficiencyhours could be directly attributed to thecontractor alone, removing the changeorder hours leaves 14,000 hours actuallyworked on the base contract (including anyloss of productivity because of the changes).

24 Cost Engineering Vol. 50/No. 12 DECEMBER 2008

Figure 1Impact of Changes on Mechanical and Electrical Work [11]

Figure 2Leonards Raw Data and Corresponding Regression Lines for Mechanical andElectrical Work, Illustrated in Terms of the Construction Productivity Index PI [14]


The percent change orders is 43percent (6,000 change order hours 14,000 actual base contract hours). Theproductivity index is 0.71 (10,000 earnedhours 14,000 actual hours) and the loss ofproductivity is 29 percent. As such, 4,000hours were lost productivity out of a total of14,000 hours actually worked on the basecontract.

Although often referred to incontractor claims and in the literature, theLeonard work has not been universallyaccepted. One common criticism is that theresearch analyzed only projects that hadalready reached the dispute stage, whichquite likely resulted in loss of productivitycurves that were skewed to the moredisturbed end of the spectrum.

None of his six curves extend to PI >1for zero percent change, which is notreasonable. Other voiced objectionsinclude a small data set (only 57 projects), asmall ($4 million) average project size, andthe fact that all his projects were Canadian.

Additional criticisms of the Leonardmethod can be found under GeraldMcEniry, William Ibbs and Dr. K.M.J.Harmon and B. Cole [5, 8, 14].Nevertheless, the principal author of thepresent article has used both Leonard andIbbs data in trial, arbitration, and mediationsettings.

Ibbs Studies (1995 and 2005)Working with both owners and

contractors, Ibbs has collected data from170 projects for the past 12 years [6, 7, 8].Projects include both public and privateprojects with different project deliverysystems, and range in size from $2 millionto $14 billion with a median value of $44million.

All project data (e.g., productivity,actual project hours, change orders,contractor errors, etc.) were obtaineddirectly from one of the project principals.Though that database includes change andproductivity information for both the designand construction phases of projects, onlythe construction information will bediscussed here.

The 1995 study contained 104 projectsbut has grown over time. It reported that nomeaningful change - productivitydifferences existed for architectural / civilvs. electrical / mechanical projects, so nodistinction is made here.

Contrary to Leonards study, severalprojects were found where productivity

exceeded plan (PI > 1). On the other hand,projects affected by > 25 percent changeorders were so sparse, that the alignment ofan extrapolated trend line could bedramatically affected by a few distantpoints. The 2005 study collected manymore data points in the range of 20 -50percent change orders, as illustrated infigure 3:

An examination of the Leonard andIbbs data led McEniry to suggest that thedata might be compatible in certain rangesof percent change orders [8, 11, 14]. Inorder to address that question though, theLeonard and Ibbs studies had to bemodified so that the percent of changeorders was calculated in the same way.

For this article, Ibbss change orderpercentages have been adjusted to conformto Leonards definition (see equation 2).The following section compares the twostudies.

Leonard and Ibbs Studies ComparedBecause there was no discernable

difference between electrical/mechanicaland architectural/civil projects in the Ibbsstudy, Leonards source data was combinedinto one data set. The results are displayedin figure 4.

The combined set of data show apronounced downward sloping curve,indicating that as a projects changeincreases, labor productivity will decline.

The best-fit equation is shown to be:

PI = 1.6911 * change2 - 1.5442 * change +0.9697

(equation 3)

R2 = 0.4951, indicating a reasonable fit.

Combining Leonards two sets of dataresults in a curve that predicts that theproductivity index is never greater than 1.0.That is, for zero change, the PI = 0.9697.This is due mainly to Leonards data, as canbe seen in figure 4 where the two sets ofdata are contrasted.

In this figure, the diamonds and linearcurve represent Leonards data and thesquares and exponential curve representIbbss data. Both show a downward slopingbehavior, meaning that productivitydecreases as change increases. Ibbss datashows a sharper loss of productivity and alsoshows that for small amounts of changepositive PI values exist.

The two models give reasonably similarresults for high levels of change. Forexample, at 40 percent change, Leonardpredicts 29 percent productivity loss andIbbs, 27 percent. At 50 percent change,Leonard predicts 30 percent and Ibbs, 34percent.

The Ibbs study has a higher R2 value(0.563 vs. 0.0668 for Leonard), meaning itis a much better predictor. Indeed, the Ibbsdata by itself has a higher R2 than thecombined data set, because the Leonarddata by itself is so scattered.

The fact that Leonards results showmore negative impact for low values ofchange (< 20 percent) is understandablesince most of his projects had alreadyreached the dispute stage. The Ibbs curvedemonstrates more positive results in thisrange since data was obtained from manyprojects seemingly unaffected by changes.The fact that both studies demonstratesignificant impact for high amounts ofchange is also logical since any project

Figure 3Comparison of the Ibbs 1995 and the 2005a Data and Curves [14]

encountering 30 to 60 percent change willbe severely impacted.

Timing of ChangesThe 2005 Ibbs study also explored the

impact of changes timing [8, 9]. Illustrativeof that later study is the graph presented infigure 5. The formulas on this figurerepresent the regression equation for eachof the three conditions: late, normal, andearly change. The R2 value is a measure ofthe curves fit to the underlying projectdata; R2 = 1.0 would be a perfect fit.

The analysis shows that projects withlate change are much more disruptive toproductivity than projects where change isrecognized earlier; e.g. at 10 percentchange, the late curve has a 20 percentproductivity loss while the normal curve hasa 10 percent loss. Moreover, early andnormal projects that have small amounts ofchange (less than 4 percent) may still havePI > 1 whereas these late change projectsalways had PI < 1.

Early, normal, and late weredefined by rank ordering and dividing theprojects into thirds. As a short-handdescriptor, early projects had 50 percent oftheir change recognized by 20 percentproject complete; normal projects, 40percent complete; and late projects, 70percent.

The reader should remember that thePI is end-of-project productivity, so a projectthat suffers much late change is incurringmore disruption and loss of productivitythan meets the eye at first glance.

Leonards thesis does not report changetiming information so no direct comparisoncould be made with Ibbss timing results.Nor does Leonards data contain designphase information, so no comparisonbetween the two studies could be made forjust the design phase or the combineddesign-construction phases.

In 2005, O. Moselhi, T. Assem and El-Rayes presented a study conductedprimarily to extend the model presentedearlier by Moselhi, Charles Leonard andFazio in 1991, to include the timing effectof change orders [16, 17]. They introduceda neural network model based on theiranalysis of 33 work packages extracted fromfiles for construction projects in Canadaand the US. Contrary to the linear increasein the timing factor proposed by Hanna,Moselhi modeled the build-up andrundown of labor hours normally spent toperform the work. The model wasincorporated into a prototype softwaresystem to estimate the loss of productivity.In addition to the developed neural networkmodel, the software incorporates four othermodels including that of Hanna 1999 [3, 4].The authors report that their modelprovides more accurate estimates of changeorder impacts on productivity.Unfortunately, no mention is made in thearticle regarding how to obtain access totheir prototype software.

Other Methods to Evaluate CumulativeImpact

A number of other methods have beenproposed or revised in recent publications.

A detailed review of these methods isprovided in William Ibbs, Long D. Nguyenand Seulkee Lee [10], WilliamSchwartzkopf [19 and updates] as well as inthe AACE International (2004)Recommended Practice 25R-03 [1, 10, 19and updates]. A few words are providedherein about some of the more well knownmethods, specifically in regards to thepresent study of the Ibbs and Leonardcurves.

In 2005, the Mechanical ContractorsAssociation revised one of their populardocuments Changes, Overtime andProductivity including explanations onhow to properly use 16 factors affectinglabor productivity [15]. Although thesefactors are frequently used to estimatespecific inefficiencies such as overtime,over manning etc. Richard J. Long suggeststhat some of the MCAA factors can also beused to quantify the cumulative impact ofchanges separate from the otherinefficiencies [12]. He indicates forexample, that the MCAA factor moraleand attitude can be caused by multiplecontract changes and rework;Reassignment of workers can be causedby unexpected, excessive changes; anddilution of supervision caused whensupervision is diverted to analyses and planchange or stop and re-plan affected work.

Of all the current methods, the MCAAfactors are the easiest to use, whichundoubtedly explains their popularity.However, the arbitrary and subjectivenature of these factors undermines theircredibility. In fact, even the MCAA notes


Figure 4Comparison of the Leonard and Ibbs Data

that factors should be applied with care since the addition ofmultiple factors can lead to unreliable results.

For instance, application of severe levels of impact for all 16factors could lead to a 43 percent loss of productivity, which isprobably not credible. Nonetheless, courts and boards in the USseem to accept the use of the MCAA factors if supported by thetestimony of an experienced construction professional, who hasbecome thoroughly familiar with the details of the specific project.Dr. K.M.J. Harmon and B. Cole summarize numerous caseswhere the MCAA factors were used [5].

In 1999, A.S. Hanna published two papers on the impact ofchange orders on productivity. The first study (1999a) [4]concerned mechanical construction and the second (1999b) [3]electrical construction. These studies found that percent change,calculated as change order hours divided by estimated basecontract hours, was more significant than the percent changedetermined by Leonard and Ibbs (change order hours divided byactual base contract hours). Also, the calculation of productivitylost was based on a multi variable empirical formula not readilyappreciated by contractors.

Considering the difference in the way the percentage changeis measured and the many other variables involved in thecalculation, the results of Hannas studies cannot be easilycompared with the Ibbs and Leonard data, at least not in agraphical format. It is understood that these studies have also notbeen endorsed in the US case law or board decisions published todate, according to Dr. K.M.J. Harmon and Cole [5].

Finally, AACE International recommends that productivitylosses be quantified using project specific studies andcontemporaneous project records, particularly the measuredmile approach [1]. The authors agree wholeheartedly with thisrecommendation. In fact, the measured mile approach was usedwhere possible to establish the project specific productivity loss.

CAUSATION AND ENTITLEMENT

A large number of change orders do not guarantee thecontractor the right to a cumulative impact claim [18]. It is simplynot sufficient to label an apparent productivity loss as theconsequence of multiple changes. Many cumulative impactclaims fail because the claimants do not establish a proper causallink between the changes and the lost productivity.

Productivity can be lost for numerous other reasons for whichthe owner has no responsibility including contractorunderestimating and inefficiencies (poor planning andorganization), weather conditions, etc.

In order to put forward a claim for cumulative impact, acontractor must demonstrate that there is a causal link between thelost productivity and the changes affecting the contract work.Some ways to do this include the development of a cause andeffect matrix [12], or a measured mile analysis qualified withdetails of the changes. It is also necessary to demonstrateentitlement by proving that while individual change orders werebeing completed, the cumulative impact was either notforeseeable, not quantifiable, otherwise not included or notallowed when pricing the individual change orders.

SYNERGISTIC EFFECT

Certain recent publications by Bob McCally [13] and PatGalloway [2] refer to a synergistic effect of changes with respectto cumulative impact. They make reference to the 1990sdefinition as set forth in the Construction Industry InstitutesQuantifying the Cumulative Impact of Change Orders forElectrical and Mechanical Contractors which offers thisdefinition:

The theory of cumulative impact claims holds that thecontractor fails to foresee the synergistic effect of changes on the


Figure 5Timing of Changes [8]


work as a whole when pricing individual changes, and therebyreceives less than full compensation.

According to this theory of cumulative impact, the issuance ofa multitude of change orders creates disruption that exceeds thedisruption caused by the individual change orders when viewedindependently. In this sense, cumulative impact would besynergistic. A contractor cannot reasonably be expected to foreseea synergistic effect when it cannot foresee the number or size of thechanges to come.

McCally [13] furthermore indicates that only theunforeseeable and unquantifiable impacts can be considered aspure cumulative impact. While this limited definition ofcumulative impact may be the ultimate goal for researchers andclaims analysts to establish, the fact is that it is very difficult toisolate this limited synergistic effect.

Although the language of most contract clauses today requiresthat the contractor include all direct and indirect costs in theapproved change order, in fact this is rarely the case. Many ownersrefuse to include the estimated cost of the impacts in changeorders and many contractors reserve their right to claim impactcosts at a later time when such impacts are known. As a result, thetotal approved change order hours reflect only direct hours andrarely include the individual impactseven though some of theimpacts may have been foreseeable and quantifiable at the time ofthe change order.

At the end of the project the contractor may be confrontedwith the situation of having a large number of excess hours workedthat cannot be explained by approved change orders or evencontractor inefficiencies (if admitted). The unexplained hours arepooled into what is more commonly labeled today as acumulative impact claim.

In addition to the pure synergistic effect, the more commoncumulative impact claim of today might include underestimateddirect work hours and foreseeable impacts hours that could have,or should have been allocated to individual changes and removedfrom the pool. There is also a possibility that pooled hours mightinclude hours for disputed changes (i.e., not in the approvedchange order total) that should also have been removed from theclaim and treated separately.

In this research, both Leonard and Ibbs have attempted toremove extraneous hours related to contractor inefficiency,underestimating and disputed changes from their calculation oflost productivity, wherever these hours were identifiable.

Best efforts have been made to establish an appropriatemeasure of cumulative impact. It was however necessary to assumethat approved change order hours include foreseeable andquantifiable impacts for individual changes (if any).

C umulative impact is not just a theoretical concept but areal occurrence on construction projects sufferingnumerous changes, the impact of which is difficult torecognize as individual change orders are issued and priced. Thisis a result of difficulties in quantifying and pricing the impacts andalso resistance on the part of owners to recognize such impacts inthe change order.

This article combined the work of two notable studies by Ibbsand Leonard to quantitatively measure the impact project changeactually had on construction labor productivity. The results of thiscomparison clearly demonstrate that increasing amounts of project

change will have significant and progressively worsening impacton labor productivity.

The curves presented in this article can be used as guidelinesin evaluating the cumulative impact of multiple changes on laborproductivity. The curves stem from actual project records, areimpartial and are easy to use. Though not perfect, the trends seemlogical. These curves should not be considered a completesolution to the dilemma of evaluating the cumulative impact ofchanges. Research is constantly evolving, and more cases shouldbe analysed and discussed to confirm the results. Alternativemethods should be used to corroborate any results.

Naturally it is best if change can be identified, measured, andevaluated on a case-by-case basis, according to Ibbs, Nguyen, Lee[10]. In those situations the impact is easier to evaluate andnegotiate, whether it is through a time-and-material or forwardpricing mechanism. Most importantly, the possible impact shouldnot be ignored. However, when many changes start to accumulateand even interact with each, evaluation and resolution are moredifficult. Generally, a measured mile approach is preferred, butthere are many times when no measured mile can be obtained.

In such cases, an approach like that presented in this articlemay be helpful. Though specificity is always preferred, generalindustry statistics can be corroborative in the hands of an impartialchange expert. Just like generalized test scores are used to admitstudents to college (SATs) or law school (LSATs), generalizedindustry statistics like the Leonard and Ibbs curves have value inestimating changes impact on labor productivity.

Of course it is important to use models like these judiciously.For instance, a contractors actual project experiences must beadjusted for bid mistakes and other inefficiencies. In addition,hours associated with disputed changes, or foreseeable impactsthat could have been allocated to individual changes, should beremoved and treated separate from the cumulative impact.Causation and liability of change must also be fully demonstratedin order to be compensated.

Still, if proper adjustments are made and these studies areapplied impartially, information like that presented herein can bequite useful, whether predicatively with each change order in aneffort to reduce cumulative impact, or later in a more globalfashion to resolve disputes. The fact that the principal author ofthis article has used his study in actual litigation demonstrates thatthis type of research can be useful to the construction industry.

REFERENCES1. AACE International Recommended Practices in Estimating

Lost Labor Productivity in Construction Claims. AACEInternational RP 25R-03, (2004).

2. Galloway, P. Cumulative Impact, Nielsen WursterCommuniqu, Volume 2.6, June 2007.

3. Hanna, A. S., J. S. Russell, E.V. Nordheim, and M.J.Bruggink. Impact of Change Orders on Labor Efficiency forElectrical Construction, Journal of ConstructionEngineering and Management, ASCE 125, 4 (1999b), 224-232.

4. Hanna, A.S., J.S. Russell, T.W. Gotzion, and E.V. Nordheim.Impact of Change Orders on Labor Efficiency for MechanicalConstruction, Journal of Construction Engineering andManagement, ASCE 125, 3 1999a), 176-184.


5. Harmon, Dr. K.M.J., B. Cole. Loss of Productivity Studies -Current Uses and Misuses : parts 1 and 2, ConstructionBriefings, August and September 2006.

6. Ibbs, C. William. Quantitative Impacts of Project Change:Size Issues, ASCE Journal of Construction Engineering andManagement. 123, 3 (1997), 308-311.

7. Ibbs, C. William and Walter E. Allen. Quantitative Impacts ofProject Change, Source Document 108, ConstructionIndustry Institute, University of Texas at Austin, Austin, Texas,(1995).

8. Ibbs, William. Impact of Changes Timing on LaborProductivity, Journal of Construction Engineering andManagement, ASCE 131, 11 (2005a), 1219-1223.

9. Ibbs, William and Min Liu. Improved Measured Mile AnalysisTechnique, Journal of Construction Engineering andManagement. 131, 12 (2005a), 1249-1256.

10. Ibbs, William, Long D. Nguyen, and Seulkee Lee. QuantifiedImpacts of Project Change, Journal of Professional Issues inEngineering Education and Practice, (2007): 45-52.

11. Leonard, Charles A. The Effects of Change Orders onProductivity, M.S. Thesis, Concordia University; Montreal,Quebec, (1988).

12. Long, Richard, J. Cumulative Impact Claims, available atwww.long-intl.com, (2005).

13. McCally, Bob, The Cumulative Impact of Change, 2006AACE International Transactions, CSC.05.

14. McEniry, Gerald. The Cumulative Effect of Change Orders onLabor Productivity - the Leonard Study Reloaded, The RevayReport, 26, 1 (2007) (available at www.revay.com).

15. Mechanical Contractors Association of America.Management Methods Committee, Bulletin No. 58, Jan,1979, revised in 1994 and 2005 as PD-2 Factors AffectingLabor Productivity. (1994).

16. Moselhi, O. Assem, I., El-Rayes, Change Orders Impact onLabor Productivity, ASCE JCEM (March 2005): 354-359.

17. Moselhi, O., C. Leonard, and P. Fazio. Impact of ChangeOrders on Construction Productivity, Canadian Journal ofCivil Engineering, 18, 3 (1991): 484-492.

18. Pittman Construction Company; GSBCA No. 4923, No.4897; 81-1 BCA P 14847; Dec. 24, 1980.

19. Schwartzkopf, William. Calculating Lost Labor Productivityin Construction Claims, Wiley, New York, (1995).

ABOUT THE AUTHORSDr. William Ibbs, M.ASCE, is a professor of Construction

Management with the Department of Civil and EnviromentalEngineering at the University of California at Berkeley. He is alsopresident of The Ibbs Consulting Group of Oakland, Calif. Hecan be contacted by sending e-mail to: [email protected]

Gerald McEniry, CFCC PSP, is a senior consultant withRevay and Associates of Montreal, Quebec, Canada. He can becontacted by sending e-mail to: [email protected].

evaluating the cumulative impact of changes on labour productivity

Documents