Anda di halaman 1dari 14

624 IEEE TRANSACTIONS ON SOFTWARE ENGINEERING, VOL. 35, NO.

5, SEPTEMBER/OCTOBER 2009

Impact of Budget and Schedule Pressure on


Software Development Cycle Time and Effort
Ning Nan and Donald E. Harter, Member, IEEE

Abstract—As excessive budget and schedule compression becomes the norm in today’s software industry, an understanding of its
impact on software development performance is crucial for effective management strategies. Previous software engineering research
has implied a nonlinear impact of schedule pressure on software development outcomes. Borrowing insights from organizational
studies, we formalize the effects of budget and schedule pressure on software cycle time and effort as U-shaped functions. The
research models were empirically tested with data from a $25 billion/year international technology firm, where estimation bias is
consciously minimized and potential confounding variables are properly tracked. We found that controlling for software process, size,
complexity, and conformance quality, budget pressure, a less researched construct, has significant U-shaped relationships with
development cycle time and development effort. On the other hand, contrary to our prediction, schedule pressure did not display
significant nonlinear impact on development outcomes. A further exploration of the sampled projects revealed that the involvement of
clients in the software development might have “eroded” the potential benefits of schedule pressure. This study indicates the
importance of budget pressure in software development. Meanwhile, it implies that achieving the potential positive effect of schedule
pressure requires cooperation between clients and software development teams.

Index Terms—Cost estimation, time estimation, schedule and organizational issues, systems development.

1 INTRODUCTION

R APID advances in technology have enabled more and


more individuals and companies to compete in the
global marketplace [1], [2], [3]. To succeed, companies must
In activities such as sales or sawmill operation, organiza-
tion researchers have long recognized a potential positive
effect of job-related pressure on task performance [9], [10],
constantly look for new ways to act faster and incur less cost [11], [12], [13], [14]. Compared with work examined by
than other players in the business. Software companies are organizational studies, software development is different in
not exceptions. Globalization, outsourcing and open-source that it tends to have higher level of uncertainty and
movements have intensified the competition in the software variability [15]. The distinct characteristics may confound
industry [4], [5], [6], [7]. To succeed, software companies the effect of pressure on performance. For example, knowing
strive to get their jobs done faster for less money. that management might reduce a software development
Consequently, schedule and budget compression has team’s estimate, a team could increase its estimate by a safety
become the norm in today’s software industry [8]. factor (i.e., “padding”) [8]. Such practice is enabled, and
In this study, we attempt to understand how the pressure maybe encouraged, by the uncertainty of software develop-
created by budget and schedule compression affects the ment. The estimation bias diminishes effects of schedule or
actual cycle time and cost of software development. budget pressure because it creates artificial resource com-
Particularly, we intend to see whether improved software pression without actually increasing the task demand. In
schedule and cost can be achieved with proper management
addition, product size, process maturity, and software
of the budget and schedule pressure. An understanding of
complexity may vary substantially across different projects.
when and how budget and schedule pressure will positively
The variability of software development has caused incon-
(or negatively) affect cycle time and cost of software
sistent results about the effect of schedule pressure on project
development projects enables managers to rationalize their
cost (e.g., among [16], [17], [18], [19]). Given the special
budget and schedule setting policies. More importantly,
characteristics of software development, a study with proper
software management can leverage the beneficial effect of
control of the confounding variables is needed to evaluate
budget or schedule pressure on software development to
the applicability of the positive effect of pressure identified in
gain more competitive advantage in the global market.
the organizational studies to software management.
Furthermore, our concern with effects of pressure is
. N. Nan is with the Price College of Business, University of Oklahoma, motivated by the uneven research attention paid to schedule
307 West Brooks, Norman, OK 73019. E-mail: nnan@ou.edu. versus budget pressure. Compared with the literature on
. D.E. Harter is with the Whitman School of Management, Syracuse
University, 721 University Avenue, Syracuse, NY 13244.
schedule pressure, budget pressure has received little
E-mail: dharter@syr.edu. attention. Extant studies often view budget as a fixed value
Manuscript received 28 July 2007; revised 3 July 2008; accepted 17 Feb. 2009; (e.g., [20] and [21]). However, it’s very likely for software
published online 20 Mar. 2009. clients to reduce a vendor’s estimated budget during contract
Recommended for acceptance by D. Rombach. negotiation. This warrants a study looking at effects of
For information on obtaining reprints of this article, please send e-mail to:
tse@computer.org, and reference IEEECS Log Number TSE-2007-07-0229. budget pressure to identify the beneficial or excessive range
Digital Object Identifier no. 10.1109/TSE.2009.18. of budget compression.
0098-5589/09/$25.00 ß 2009 IEEE Published by the IEEE Computer Society
Authorized licensed use limited to: Sreenidhi Institute of Science & Technology. Downloaded on November 5, 2009 at 04:49 from IEEE Xplore. Restrictions apply.
NAN AND HARTER: IMPACT OF BUDGET AND SCHEDULE PRESSURE ON SOFTWARE DEVELOPMENT CYCLE TIME AND EFFORT 625

By borrowing insights from organizational studies, we Jeffery [16], these early studies concerned the sensitivity of
propose nonlinear relationships between budget and project cost to variation in overall elapsed development
schedule pressure and development outcomes. With data time (including both schedule compression and expansion).
collected on software projects from a $25 billion/year Although they disagreed on the effect of schedule expan-
international technology firm, this study empirically tests sion on project cost, all three studies predicted “increased
the relationship between budget and schedule pressure and total effort when the manager’s constraint on elapsed time
project performance. The data set had adequate information available for development is compressed below the ‘opti-
to control for several critical confounding variables, such as mum’” [16, p. 852].
the estimation bias, product size, process maturity, software Subsequent studies in the literature offered more insights
complexity, and conformance quality. into the effect of schedule compression on software develop-
The scope of the current study is described by the ment. For example, Jeffery [16] questioned the general-
following concepts and research questions. Budget and izability of the negative effect of schedule compression
schedule pressure is defined by the discrepancy between predicted by earlier studies. By analyzing data from
budget and schedule estimated by the development team commercial software projects in Australia, he found that a
and those negotiated with the client. Software development compressed schedule could lead to either increased or
refers to all the stages starting from initial design through decreased development effort. Following Jeffery’s work,
final product acceptance testing. Based on the tradition of Kitchenham [23] found similar results by analyzing another
software engineering research, the cost of software devel- data set from a European software project. Moreover, by
opment is measured by the total number of person months reanalyzing the data set behind COCOMO, she found that
logged by the development team [22] and is referred to as there were schedule-compressed projects that were also
development effort throughout this paper. Cycle time, or effort compressed, although COCOMO predicted a negative
project duration, is measured by the calendar months relationship between the two. These studies attributed the
elapsed from the first day of design to final customer contradictory findings to difference among research sites,
acceptance of the product. Our research question is: What is variation among measures of schedule pressure, and missing
the impact of budget and schedule pressure on software control variables in the relationship between schedule
development cycle time and effort controlling for software pressure and development effort [16], [23]. In general, these
process, size, complexity, and conformance quality? subsequent studies enriched our understanding of the
The rest of the paper is organized as follows: In the next relationship between schedule pressure and development
section, we review the software engineering literature about
outcomes. Empirically, we see that reduced schedule is not
the pressure-performance relationship. In Section 3, we
always achieved at the expense of increased effort; a proper
develop the theoretical foundation and hypotheses of this
level of schedule pressure may have beneficial effects on
study. Section 4 describes our research method. Section 5
project outcomes.
presents the statistical analysis and results. Section 6
In recent years, a few studies have tried to explain the
examines the results. In the last section, we discuss
underlying mechanisms of the impact of schedule pressure.
conclusions of the study.
For example, Austin [24] employed an agency framework to
characterize the strategic behavior of software developers
2 LITERATURE REVIEW under schedule pressure. He reduced the concern of soft-
In the software engineering literature, researchers have long ware developers to two main aspects: 1) being singled out as
recognized the critical impact of resource constraints (e.g., the only one who cannot finish task on time and 2) being
project schedule or budget) on software development penalized for taking quality compromising shortcuts.
outcomes. Particularly, with respect to project schedule, a Results derived from the mathematical model indicate that
stream of research has shed light on its effects on contrary to casual intuition, schedule pressure could
development cycle time or effort. For example, in The maintain or even improve software project performance.
Mythical Man-Month, Brooks pointed out that schedule and The underlying mechanism was that software developers
manpower were not interchangeable because software took quality impairing shortcuts to avoid being singled out
developer productivity may deviate significantly from a as the only few who could not meet deadlines. Very tight
constant rate as complexity of software projects increases. schedules made it impossible for anyone to meet deadlines.
He stated the viewpoint as Brooks’ Law: “adding man- When software developers did not worry about being
power to a late software project makes it later” [19]. With singled out, they would be less likely to take shortcuts and
Brooks’ Law as the underlying assumption, Putnam concern more about quality improvement. Higher quality
proposed a trade-off function between time and effort in incurred less rework, and thereby, reduced development
the Rayleigh curve model [17]. The function indicates that cycle time and effort [25], [26], [27].
any compression of schedule has to be compensated by an Using a system dynamics model, Abdel-Hamid et al.
increase of effort to maintain the same project cost. [28], [29], [30], [31] proposed multiple causal pathways
Similarly, the COnstructive COst MOdel (COCOMO) [18] between schedule pressure and software development
predicted that there existed an optimal schedule value; any outcomes. They recognized that schedule pressure not only
compression or lengthening from the optimum would lead had a direct impact on productivity by driving developers
to higher cost. Together, Brook’s Law, Putnam’s Rayleigh to work long hours but also produced an indirect impact by
curve model, and COCOMO represent the early “trade-off” affecting the error generation rate of a project. Results from
view about impacts of project schedule. As summarized by the system dynamics model suggested that contrary to

Authorized licensed use limited to: Sreenidhi Institute of Science & Technology. Downloaded on November 5, 2009 at 04:49 from IEEE Xplore. Restrictions apply.
626 IEEE TRANSACTIONS ON SOFTWARE ENGINEERING, VOL. 35, NO. 5, SEPTEMBER/OCTOBER 2009

TABLE 1
Summary of the Software Literature

Brooks’ Law, adding more people to a late project did not effort. Second, due to difference in research sites and
always make it completed later; only aggressive compres- measures, empirical findings alone may be inconclusive
sion of schedule could cause higher development effort and (see the discussion in [16]). Thus, theoretical explanation of
longer cycle time. the underlying mechanisms is warranted to improve the
Compared with studies about schedule pressure, re- generalizability of previous findings. Although the recent
search on budget pressure is rather scanty in the software studies have offered some theoretical insights, there is a
literature. Budget constraint is usually discussed as a given need for more fine-grained understanding about the
condition (e.g., [20] and [21]) or a value to be minimized fundamental forces beneath the nonlinear relationship
subject to certain outcomes (e.g., [32]). The few exceptions between pressure and job performance. Third, there is a
only provide brief arguments about possible impacts of knowledge gap regarding the impact of budget pressure on
budget compression without performing any formal ana- software development outcomes. While clients commonly
lysis. For example, the gap between original estimate and reduce a development team’s proposed budget during
current budget is discussed as one of the questions for project negotiation, the impact of budget pressure has not
software feasibility review [33]. In addition, budget been sufficiently addressed in the literature.
compression is briefly mentioned as a potential cause of To formally characterize the nonlinear impact of schedule
higher development cost [34]. pressure and explore the effect of budget pressure, we borrow
The software literature is summarized in Table 1. Taken insights from organizational studies. Researchers have
together, the literature has three important implications. identified fundamental forces underlying the relationship
First, the impact of schedule constraint on project outcomes between job-related pressure and employee performance
is nonlinear. Particularly, depending on the degree of [35]. The fundamental forces are an employee’s natural
schedule compression, schedule pressure may have either responses to stress such as budget or schedule pressure. The
positive or negative effect on development cycle time and aggregation of the fundamental forces can generate a

Authorized licensed use limited to: Sreenidhi Institute of Science & Technology. Downloaded on November 5, 2009 at 04:49 from IEEE Xplore. Restrictions apply.
NAN AND HARTER: IMPACT OF BUDGET AND SCHEDULE PRESSURE ON SOFTWARE DEVELOPMENT CYCLE TIME AND EFFORT 627

nonlinear relationship between pressure and performance amount of resources (e.g., time and budget) [47]. Higher
[9], [10], [11], [36], [37], [38], [39], [40], [41], [42], [43]. budget or schedule pressure means more difficult goals
[48], [49]. A major implication of the goal setting theory is
that goal acceptance interacts with goal difficulty and leads
3 THEORETICAL DEVELOPMENT to a nonlinear relationship between goal difficulty and job
Nearly a century ago, researchers found that performance performance [44]. Specifically, goal acceptance is negatively
could increase with intensity of task demand, but only to a related to goal difficulty, with the transition from accep-
certain point [9]. When intensity of task demand became too tance to rejection usually occurring at an intermediate level
high, performance would decrease. Overall, optimal per- of goal difficulty. Goal difficulty has a positive effect on
formance tended to occur at an intermediate level of task performance for accepted goals and a negative effect on
demand while very low or very high level of task demand performance for rejected goals. Overall, performance first
often led to poor performance. This nonlinear relationship increases with goal difficulty and then levels off or
between task demand and performance is often referred to decreases when goals are too difficult to be accepted.
as Yerkes-Dodson Law. Optimal job performance usually occurs under an inter-
Organization researchers have shown the generalizabil- mediate level of pressure.
ity of Yerkes-Dodson Law to the relationship between job- Previous research has extended the pressure perspec-
related pressure (e.g., budget or schedule pressure) and tives from an individual level to a team level. For example,
employee performance [35]. For example, drawing on the numerous studies have shown that individual-level goal
theoretical framework of Yerkes-Dodson law, Singh [35] setting effects can be generalized to teams [47], [50], [51],
hypothesized curvilinear impacts of role stressors (i.e., role [52]. The resemblance between individual and team-level
conflict, ambiguity, and overload) on salespeople job pressure effects is most evident when teams are temporarily
outcomes such as performance, job tension, and satisfac- assembled and each member’s expected and actual con-
tion. Empirical data from a range of small and large firms tribution is identifiable. Temporarily assembled teams are
supported the curvilinear effect of role stressors on exempted from confounding effects of a number of team
salespeople job tension. Meanwhile, in a field experiment contextual variables such as team history or team culture
about sawmill workers, researchers found that both work [51]. Meanwhile, visibility of individual member’s expected
underload (e.g., mechanically controlled work pact, repeti- and actual contribution prevents social loafing [53], which
tion of short-cycle operations) and overload (e.g., piece- may cause deviation in a team’s reaction to pressure due to
rate rush and high demands of attention) could cause long- the individual members’ lack of effort [51].
term negative effects on workers’ well-being, health, and In this study, we extend the nonlinear relationship
job satisfaction. Once workers were given control of the between pressure and performance from an individual level
pace and methods of their work, they experienced a to a team level. This extension is based on two premises. First,
moderate level of pressure and showed significantly better in the software industry, an essential strategy of project
health and job satisfaction results. Moreover, in a labora- management is the contracting of IS professionals on a
tory experiment, researchers varied the level of goal temporary basis [54]. This strategy helps organizations to
difficulty in a perceptual speed test. They found that the remain responsive to economic and technological changes by
relationship between performance and goal difficulty reducing the fixed cost of permanent staff [55], [56]. It makes
changed from positive to negative as researchers raised temporarily assembled software development teams very
the goal difficulty level [44]. common in the software industry. As discussed above,
In general, the applicability of Yerkes-Dodson law is temporary teams are less affected by team contextual
based on the understanding of the mechanisms behind the variables and, therefore, are more likely to show a nonlinear
nonlinear relationship between pressure and performance. pressure-performance relationship similar to that on an
Researchers have realized that employees’ fundamental ego- individual level. Second, the basic methodology of software
needs, such as challenge and pride in their work, mediate the programming is divide-and-conquer, that is, dividing the
pressure-performance relationship [14]. They see that a low overall task into parts and working on each part individually
level of job-related pressure often means a lack of challenge. [57]. The divide-and-conquer method makes each team
When employees (e.g., software developers) cannot fulfill member’s expected and actual contribution identifiable.
their fundamental need for challenge, they tend to be Thus, the confounding effect of social loafing on the
inattentive and bored and, thus, do not perform well [45], pressure-performance relationship is minimized.
[46]. On the other hand, a high level of pressure often creates In summary, the organizational studies provide fine-
too much challenge for employees to handle. Their pride in grained understanding of mechanisms underlying the
work may be threatened [14]. Consequently, they tend to feel nonlinear relationship between pressure and performance.
anxious and frustrated and, hence, perform poorly. Only They have indicated the generalizability of the nonlinear
under an intermediate level of pressure, employees are more relationship to studies about effects of continuous task
likely to fulfill both needs for challenge and pride in work. demand, such as budget or schedule pressure, on project
Optimal performance usually follows the fulfillment of team performance.
fundamental ego-needs.
Goal setting theory (see [47] for a review) implies another 3.1 Hypotheses
mechanism for the nonlinear relationship between pressure In the software development context, better performance is
and job performance. A goal is the object or aim of a job indicated by shorter cycle time and less development effort.
(e.g., to complete a software product) with a specified Based on the nonlinear view, shortest cycle time and lowest

Authorized licensed use limited to: Sreenidhi Institute of Science & Technology. Downloaded on November 5, 2009 at 04:49 from IEEE Xplore. Restrictions apply.
628 IEEE TRANSACTIONS ON SOFTWARE ENGINEERING, VOL. 35, NO. 5, SEPTEMBER/OCTOBER 2009

TABLE 2
Budget and Schedule Pressure Hypotheses

development effort tend to occur under an intermediate developers and managers [35]. None of these measures
level of budget or schedule pressure while longer cycle time fulfills the three requirements discussed above. The metrics
and higher development effort are associated with very low based on actual elapsed time [16], [17], [18] are post hoc
or very high level of pressure. Overall, we propose rather than ex ante. They do not indicate the pressure level
U-shaped relationships between budget/schedule pressure prior to performance. The self-reported values [35] depend
and software development cycle time and effort. With the on people’s subjective opinions. It is difficult to compare
U-shaped view between pressure and software develop- personal assessments across development teams.
Due to a lack of established measures of budget or
ment performance outcomes, we hypothesized the follow-
schedule pressure, we developed an ex ante measure of
ing answers to our research question regarding impacts of
pressure:
budget or schedule pressure on software development cycle
time and development effort (see Table 2). Budget Pressure ¼
Team-Estimated Budget  Customer-Negotiated Budget
;
4 RESEARCH METHODOLOGY Team-Estimated Budget
4.1 Construct Measures
4.1.1 Budget and Schedule Pressure Schedule Pressure ¼
In the scope of this study, we see three requirements for Team-Estimated Time  Customer-Negotiated Time
:
effective measures of budget and schedule pressure. First, Team-Estimated Time
according to their definition, proper measures should Budget refers to the number of person months of effort
indicate the discrepancy between development team-esti- allocated for a development project. Cycle time (or project
mated budget or schedule and the customer-negotiated duration) is the number of calendar months elapsed from
budget or schedule. Second, the literature on the pressure the first day of design to final customer acceptance of the
effect implies that people have to first perceive the pressure product. At the research site, development teams first
and then respond to it through their performance. As proposed their estimated budget and cycle time with the
Robert pointed out that “Software’s problems . . . are about assistance of an estimation tool. During the negotiation
expectations established at the outset of a project much between the corporation and clients, the customer usually
more than they are about trouble that happens along the reduced team-estimated budget, with an average reduction
way” [58, p. 19]. Therefore, measures should be taken at the of 9.5 percent but never an increase in the budget. The
onset of a project rather than at the completion of it. Third, customer was more flexible in negotiations on schedule.
they should be quantitatively comparable across projects The average schedule reduction was 11.6 percent, but with
and teams. This requires a formula that can standardize and both expansion and compression of schedules occurring in
quantify the pressure level of a development team. negotiations. The customer-reduced budget and modified
The literature has not established a consistent measure cycle time are recorded in the final contracts after the
for budget or schedule pressure. Budget pressure, often completion of negotiation with the client. The ratio of the
referred to as “budget constraint” or “cost constraint,” is change to the budget or schedule to the original budget or
usually measured as the total amount of capital available to schedule is the measure of pressure. For example, if a team-
a project (e.g., [18]). This measure does not reflect the estimated budget is $10;000 but the negotiation reduces the
discrepancy between team-estimated budget and customer- budget to $8;000, the budget pressure is 20 percent
negotiated budget. Moreover, it is difficult to compare (($10;000  $8;000Þ=$10;000 ¼ 0:2 or 20 percent). The nu-
budget constraints across projects. merators of the two measures reflect the discrepancy
Schedule pressure has been assessed by the elapsed time between team estimates and customer-negotiated budget
of a project (e.g., [17], [18]), the ratio between actual elapsed or cycle time. We use team-estimated budget or cycle time
time and estimated elapsed time (e.g., [16]), the ratio as the denominator because taking the ratio between the
between scheduled completion date and forecasted com- discrepancy and the initial team estimates can quantify and
pletion date [30], or self-reported values by software standardize pressure levels across projects and teams.

Authorized licensed use limited to: Sreenidhi Institute of Science & Technology. Downloaded on November 5, 2009 at 04:49 from IEEE Xplore. Restrictions apply.
NAN AND HARTER: IMPACT OF BUDGET AND SCHEDULE PRESSURE ON SOFTWARE DEVELOPMENT CYCLE TIME AND EFFORT 629

Precisely speaking, our measures of budget and assurance, defect prevention, peer review, and training [68].
schedule pressure only capture the postestimation budget The more CMM process areas that are adopted, the higher
and schedule compression. They do not account for budget the maturity level is.
and schedule pressure that arises and evolves during a Product complexity captures three important dimensions
software development project. Expectations established at of complexity: domain, data, and decision complexity [66]
the beginning of a project can be a significant source of (see the Appendix, which can be found on the Computer
variations in software development teams’ performance Society Digital Library at http://doi.ieeecomputersociety.
during a project [58]. In this paper, we focus on the org/10.1109/TSE.2009.18, for details). Domain complexity is
postestimation budget and schedule compression while the level of functional complexity as determined by engineer-
acknowledging that the pressure occurring during a project ing from the user requirements specification. Data complex-
can provide a different perspective of the relationship ity refers to the anticipated level of difficulty in developing
between pressure and performance. the system because of complicated data structures and
database relationships. Decision complexity measures the
4.1.2 Cycle Time degree of difficulty in the decision paths and structures
Cycle time is one dimension of the project performance. It is within the product. Software teams assessed the three
the time to develop the software product, i.e., the number of dimensions of product complexity using the criteria specified
calendar months elapsed from the first day of design to in the Appendix, which can be found on the Computer
final customer acceptance of the product. Society Digital Library at http://doi.ieeecomputersociety.
org/10.1109/TSE.2009.18. Each software product manager
4.1.3 Development Effort
was trained in the assessment of complexity by senior
Development effort is another dimension of project perfor-
management used in the estimation model [66]. The three
mance. In this study, the total number of person months
dimensions of complexity were assessed ex ante by the
logged by the development team from initial design
software product manager, monitored by quality assurance
through final product acceptance testing constitutes the
to ensure compliance with the estimation process, verified by
development effort.
senior management for consistency with other projects, and
4.1.4 Control Variables validated by an independent assessment of the client’s
technical staff and managers. The three scales of complexity
Software conformance quality is defined as the extent to
which a software product meets quality standard and initial showed high intercorrelation. Scores of the product complex-
customer specifications. Prior studies have found that ity were computed by taking an average of the three
software conformance quality, usually measured by the complexity scales.
number of defects, affects the amount of rework. Subse- 4.2 Regression Models
quently, the amount of rework has a significant impact on
We developed four research models to test the effects of
development cycle time and effort [25], [26]. Given the
budget or schedule pressure on development cycle time
importance of conformance quality in development perfor-
and effort of software projects. Their function forms are
mance, we controlled for its effect in the data analysis. Here,
the following:
conformance quality is measured as the number of lines of
source code in the product divided by the sum of defects CycleTime ¼ fðprocess maturity; size; complexity;
found in system testing. Three stages of system testing using ðaÞ
quality; budget pressureÞ;
a formal test plan were contractually mandated for all
products: vendor’s system testing, client system testing with
a sample database, and client stress testing with a production Effort ¼ fðprocess maturity; size; complexity;
ðbÞ
database. Vendor testing was performed by a test group quality; budget pressureÞ;
independent of the development team. Client testing was
performed by the client’s technical and user community,
CycleTime ¼ fðprocess maturity; size; complexity;
supported by a contractor with expertise in testing complex ðcÞ
systems. This measure of quality is the inverse of the defect quality; schedule pressureÞ;
density [59] used in many previous quality studies. This
paper adopts the inverse defect density because it offers a Effort ¼ fðprocess maturity; size; complexity;
more intuitive understanding of the conformance quality ðdÞ
quality; schedule pressureÞ:
values: Higher values mean better conformance quality.
Product size has been recognized as a significant factor 4.3 Data Collection
in cycle time [60], [61] and development effort [62], [63]. In To test the hypotheses, we collected cross-sectional data on
this study, product size is measured by thousand lines of software projects performed by a $25 billion/year interna-
source code in a product. All projects in this study used the tional technology firm. The company contracts commercial,
same language, ensuring comparability. international, and government clients. The software pro-
Past research indicated that process maturity [64], [65] jects in our data set were developed for a material
and product complexity [66], [67] have significant impacts requirements planning information system to manage spare
on software development performance. Process maturity is parts acquisition.
measured by the SEI’s CMM level of maturity. The CMM In the technology firm, assessment of process maturity
framework includes 18 key process areas such as quality was performed by government auditors and corporate

Authorized licensed use limited to: Sreenidhi Institute of Science & Technology. Downloaded on November 5, 2009 at 04:49 from IEEE Xplore. Restrictions apply.
630 IEEE TRANSACTIONS ON SOFTWARE ENGINEERING, VOL. 35, NO. 5, SEPTEMBER/OCTOBER 2009

personnel in divisions outside of systems integration. These Agrawal and Chari [67] reviewed 18 published software
independent groups used the SEI’s CMM [68] to assess the studies and found a median sample size of 37 (see the study
maturity of the firm’s software development and support- for more detail).
ing activities. We collected process maturity data from these
independent groups.
Other data were collected by the technology firm and 5 ANALYSIS AND RESULTS
were audited by the clients to ensure accuracy and To test the existence of the U-shaped pressure-performance
accountability. Cycle time data were retrieved from the relationship, we included quadratic terms of budget or
project management scheduling system. Effort data were schedule pressure as independent variables in the statistical
tracked by the corporate time reporting system. Software models. This method has been proved effective by previous
defect data were extracted from the Configuration Manage- studies (e.g., [70]). Researchers have found economies and
ment (CM) database. Information related to the estimated, diseconomies of scale [61] in software development. As a
budgeted, and actual duration and cost of software result, a log-log relationship has generally been used to study
development was manually coded from the electronic and software effort and cycle time. The appropriateness of a log-
paper files maintained by the firm. log model to our data set is further supported by the
The technology firm provided an ideal site for this Davidson and MacKinnon’s specification test for non-nested
research because its estimation method and management models (J-test) [71]. The J-test result shows that a log-log
strategies can help us eliminate alternative explanations of model is preferred over semilog and linear models. In the log-
the impacts of budget or schedule pressure on development log model, we did not take log-transformation of the pressure
performance. First, by using the Software Productivity terms because it is possible for budget or schedule pressure
Quality and Reliability (SPQR20) automated tool, the firm values to be 0 or negative. Moreover, since our models focus
on the U-shaped relationship between pressure, effort and
minimized the “padding” effect. In software industry,
cycle time, we operationally define the U-shaped function
development teams often extend their “rational” estimates
using a quadratic term. In this formulation, taking the
of cycle time or cost by a safety factor to counteract the high
logarithm of a quadratic would simply reduce it to twice
levels of pressure [8]. The magnitude of this padding effect
the linear term. The specific regression models for budget
varies across project teams and is hard to assess. As a result,
pressure analysis are the following:
many earlier research findings were confounded by the
padding effect. At our research site, development teams ln ðCycle T imeÞ ¼ 01 þ 11 ln ðprocess maturityÞ
estimated function point counts and complexity based on
þ 21 ln ðproduct sizeÞ þ 31 ln ðcomplexityÞ
user task requirement specification. These estimates were ð1Þ
validated by the client to ensure accurate interpretation of þ 41 ln ðconformance qualityÞ
the requirements. The function point counts, complexity, þ 51 ðbudget pressureÞ þ 61 ðbudget pressureÞ2 ;
and environmental variables were submitted to SPQR20 for
budget and schedule projection. SPQR20 provides a
ln ðEffortÞ ¼ 02 þ 12 ln ðprocess maturityÞ
consistent methodology for estimation of budget and
schedule, preventing distortion of the estimates. þ 22 ln ðproduct sizeÞ þ 32 ln ðcomplexityÞ
ð2Þ
To ensure estimation accuracy with SPQR20, develop- þ 42 ln ðconformance qualityÞ
ment teams used historical data to calibrate budget and þ 52 ðbudget pressureÞ þ 62 ðbudget pressureÞ2 :
schedule projections. Senior management examined the
accuracy of the methodology annually prior to estimation The statistical models for schedule pressure analysis are
of the next year’s product development. Historically, the the following:
estimation tool underestimated budget by 6.2 percent, i.e.,
the teams averaged 6.2 percent budget overruns compared ln ðCycle T imeÞ ¼ 03 þ 13 lnðprocess maturityÞ
to estimates. The calibration of SPQR20 with the organiza-
tion’s historical data ensured that development team þ 23 ln ðproduct sizeÞ þ 33 ln ðcomplexityÞ
ð3Þ
productivity was accurately reflected in the model estimates. þ 43 lnðconformance qualityÞ
Furthermore, the technology firm implemented a series þ 53 ðschedule pressureÞ þ 63 ðschedule pressureÞ2 ;
of management practices to prevent bias in software
schedule and cost estimates. These practices corresponded
to the bias reduction strategies identified by Peeters and ln ðEffortÞ ¼ 04 þ 14 ln ðprocess maturityÞ
Dewey [69]. Each of these practices was performed in the þ 24 ln ðproduct sizeÞ þ 34 ln ðcomplexityÞ
estimation of schedule and budget by the firm during the ð4Þ
þ 44 ln ðconformance qualityÞ
time frame of the study. Table 3 summarizes the strategies
recommended by Peeters and Dewey [69] and how the þ 54 ðschedule pressureÞ þ 64 ðschedule pressureÞ2 :
technology firm implemented the corresponding practices.
We found 66 projects with suitable data for the study of The models were initially tested with ordinary least
budget and schedule pressure (see Table 4 for the descriptive squares (OLS). However, because data in this study are all
statistics, and the Appendix, which can be found on the from the same company, it may be possible that the error
Computer Society Digital Library at http://doi.ieeecompu- terms are correlated as a result of some common effect. A
tersociety.org/10.1109/TSE.2009.18, for the correlation ma- Breusch-Pagan test of independence indicated that the error
trix). The sample size of our study is comparable to previous terms are significantly correlated for both the budget pressure
studies about software development. In a recent study, equations ((1) and (2)) (2 ¼ 5:29; p ¼ 0:02) and the schedule

Authorized licensed use limited to: Sreenidhi Institute of Science & Technology. Downloaded on November 5, 2009 at 04:49 from IEEE Xplore. Restrictions apply.
NAN AND HARTER: IMPACT OF BUDGET AND SCHEDULE PRESSURE ON SOFTWARE DEVELOPMENT CYCLE TIME AND EFFORT 631

TABLE 3
Bias Reduction Strategies

TABLE 4
Descriptive Statistics

pressure equations ((3) and (4)) (2 ¼ 14:96; p ¼ 0:00). indicated that no outlier existed in the schedule pressure
Therefore, we estimated the seemingly unrelated regres- data and four outliers existed in the budget pressure data.
sion (SURE) parameters using a feasible generalized least To ensure the validity of the test results, we removed the
squares (FGLS). Tables 5 and 6 display SURE estimates for outliers and performed statistical tests on the budget and
the budget pressure models ((1) and (2)) and the schedule schedule pressure data sample. The normality assumption
pressure models ((3) and (4)), respectively. was not rejected using the Kolmogorov-Smirnov test [73].
We performed Cook’s Distance test [72] with 1 as the No evidence was found of heteroskedasticity using White’s
threshold to look for outliers and influential points in both [74] test. Multicollinearity of explanatory variables was
budget and schedule pressure data samples. The results tested using the condition index and variance inflation

Authorized licensed use limited to: Sreenidhi Institute of Science & Technology. Downloaded on November 5, 2009 at 04:49 from IEEE Xplore. Restrictions apply.
632 IEEE TRANSACTIONS ON SOFTWARE ENGINEERING, VOL. 35, NO. 5, SEPTEMBER/OCTOBER 2009

TABLE 5
Parameter Estimates of the Budget Pressure Models

factor. All condition indexes and variance inflation factors The results do not support the U-shaped effect of
were within acceptable ranges. schedule pressure on development cycle time. Although
The results suggest strong support for our hypothesized the coefficient of the linear schedule pressure term in (3) is
U-shaped relationship between budget pressure and devel- negative and significant (53 ¼ 0:43; p ¼ 0:01), the coeffi-
opment cycle time. As shown in Table 5, there is a positive cient of the quadratic term in the cycle time equation (3) is
and significant coefficient (61 ¼ 11:38; p ¼ 0:05) for the not significant (63 ¼ 0:25; p ¼ 0:40) (Table 6). The results
quadratic term and a negative and significant coefficient do not support H3 that predicts schedule pressure has a
(51 ¼ 3:80; p ¼ 0:02) for the linear term of budget pres- significant U-shaped effect on software development cycle
sure in the cycle time equation (1). The statistical results time.
indicate that budget pressure has a significant U-shaped Finally, we did not find statistical support to the
effect on development cycle time. Thus, H1 is supported. hypothesized U-shaped relationship between schedule
Statistical results also show strong support for our pressure and development effort. As shown in Table 6,
hypothesized U-shaped relationship between budget pres- the coefficient of the quadratic term in the effort equation
(4) is negative but not significant (64 ¼ 0:87; p ¼ 0:06).
sure and development effort. In Table 5, the coefficient for
Also, the linear pressure term in (4) is negative and not
the quadratic term in the effort equation (2) was positive and
significant (54 ¼ 0:43; p ¼ 0:07). The results do not sup-
statistically significant (62 ¼ 31:39; p ¼ 0:01). Meanwhile,
port the U-shaped curve for H4.
the coefficient for the linear budget pressure term in the In summary, we see that budget pressure has sig-
same equation was negative and significant (52 ¼ 12:23; nificant U-shaped relationships with software develop-
p ¼ 0:00). These results suggest that budget pressure has a ment cycle time and development effort, controlling for
significant U-shaped effect on development effort. Hence, software process, complexity, and conformance quality. In
H2 is supported. contrast, schedule pressure does not show a significant

Authorized licensed use limited to: Sreenidhi Institute of Science & Technology. Downloaded on November 5, 2009 at 04:49 from IEEE Xplore. Restrictions apply.
NAN AND HARTER: IMPACT OF BUDGET AND SCHEDULE PRESSURE ON SOFTWARE DEVELOPMENT CYCLE TIME AND EFFORT 633

TABLE 6
Parameter Estimates of the Schedule Pressure Models

U-shaped relationship with development cycle time or the sampled projects. An interesting finding is that user
development effort. involvement in the software development process might
have attenuated the improvement of development cycle time
in response to schedule pressure. Specifically, the clients of
6 DISCUSSION the projects were involved in the development process by
To gain intuition of the nonlinear effect of budget pressure, conducting periodic reviews of the deliverables (e.g., design
we plot the components of the linear and squared budget specifications, test plans, and user manuals), formal life cycle
pressure terms and their combined effect on log-trans- phase reviews (design, detailed design, code development,
formed cycle time and effort (see Figs. 1 and 2). The linear and testing), and audits. This review and audit time ranged
representations of cycle time and effort appear in Figs. 3 and from 10 to 20 percent of the product development schedule.
4. As indicated in the figures, the negative linear term When comparing the client review time specified in the
initially reduces the cycle time and cost; the positive squared estimates with the actual length shown in the schedule
pressure term eventually offsets any time and cost reduction charts, we noticed that clients never shortened their review
due to the linear term. The most significant improvement time even when the projects were under schedule pressure.
due to budget pressure occurs in the range of 0-10 percent In addition, the client reviews generally appeared on the
pressure, identified as beneficial range in the figures. The critical path for the product schedules. The lack of elasticity
combined curves tend to flatten and become shallow over of client review time reduced the variability of the develop-
10 percent, noted in the limited difference range. Excessive ment cycle time in response to schedule pressure. This
budget compression results in rapidly increasing cycle time implies that achieving the potential positive effect of
and costs, shown as the excessive compression range. schedule pressure requires cooperation between clients and
The nonsignificant effect of schedule pressure is some- software development teams.
what surprising. To further interpret the result, we examined In contrast, the lack of effect of schedule pressure on
the estimates, contracts, and detailed activity schedules of effort can be attributed to other causes. Specifically, when

Authorized licensed use limited to: Sreenidhi Institute of Science & Technology. Downloaded on November 5, 2009 at 04:49 from IEEE Xplore. Restrictions apply.
634 IEEE TRANSACTIONS ON SOFTWARE ENGINEERING, VOL. 35, NO. 5, SEPTEMBER/OCTOBER 2009

Fig. 1. Linear and squared component effects of budget pressure on Fig. 3. Effect of budget pressure on cycle time.
log(cycle time).
monthly against this baseline to determine if the project was
schedule pressure was not accompanied by significant tracking to the planned schedule and budget. We utilize the
budget pressure, the project manager had the flexibility of earned value charts generated during this management
increasing manpower over a shorter period of time without process as the lens to open up development teams’ reaction
affecting the budget. For example, a project planned for five to budget and schedule compression.
months with four software developers, when shortened to Figs. 5 and 6 are the earned value charts for low- and
four months, could increase staff to five software devel- high- pressure projects, respectively. In the figures, planned
opers, maintaining 20 person months of effort. The benefit value (PV) is the dollar value of work scheduled. Earned
of schedule pressure on effort was negated by the increase value (EV) is the actual value of work completed; actual cost
in staff by management. (AC) is the dollar expenditure that was used to accomplish
Investigation into whether the performance of the the work to date. The latest revised estimate (LRE) reflects
estimation tool affected the results reveals an interesting management projection of an underrun or overrun. The
pattern. For all projects, the estimation tool underestimated contract line represents what was negotiated in the contract;
projects, i.e., the average project resulted in a 6.2 percent adjustments to the contract can occur when functionality is
budget overrun compared to the original estimate. How- changed, altering the contract value.
ever, for projects with beneficial pressure, those with less Fig. 5 is an earned value chart for a project with budget
than 10 percent budget pressure, projects, on average, pressure in the 0-10 percent range. Under the lower
experienced a 7.7 percent budget underrun. Projects with pressure scenario, the AC exceeded the PV in the first
more than 10 percent budget pressure had average budget month, but the development team appeared to be motivated
overruns of 18.8 percent. by this cost overrun and able to recover in the second
Furthermore, to deepen our understanding of the month. In the sixth month, the EV lagged the PV, indicating
mechanisms underlying the observed pressure effects, we schedule slippage. Again, the team was able to correct
examined development teams’ reaction to extremely high- before the end of the project. The latest revised estimate was
and low levels of budget and schedule reductions. At the stable, a sign that the development team was able to react
research site, management used the Earned Value approach positively to any temporary schedule or cost slippage and
to track performance against the project baseline during a make adjustments throughout the life of a project to
software development project. The baseline is the planned accommodate low levels of pressure.
schedule and budget of the project from final negotiations. Under a higher pressure scenario, as seen in Fig. 6, the
The actual schedule and cost performance were monitored effect of pressure on budget and schedule resulted in cost
and schedule overruns. In the figure, it appears that there

Fig. 2. Linear and squared component effects of budget pressure on


log(effort). Fig. 4. Effect of budget pressure on effort.

Authorized licensed use limited to: Sreenidhi Institute of Science & Technology. Downloaded on November 5, 2009 at 04:49 from IEEE Xplore. Restrictions apply.
NAN AND HARTER: IMPACT OF BUDGET AND SCHEDULE PRESSURE ON SOFTWARE DEVELOPMENT CYCLE TIME AND EFFORT 635

Fig. 5. Project with beneficial pressure. Fig. 6. Project with high pressure.

were several points in time when the development team management’s ability to increase staffing when schedules
recognized that schedule and budget goals were too were reduced.
challenging to achieve, as indicated in changes in the Furthermore, when previous studies imply a nonlinear
LRE line: after the first few months and again approxi- impact of resource constraints on software development
mately 50 percent through completion of the project. The outcomes, results of this study suggest that such nonlinear
problems were exacerbated by functional changes to the impacts can be formalized as a U-shaped function. Our data
product baseline late in the life cycle at months 13 and 14, analysis indicates that budget pressure has a positive
resulting in a final spiking of the LRE in month 14. Overall, impact on software development performance when it is
in this high pressure scenario, the development team’s within a beneficial range of 0-10 percent. Beyond this
attempt to catch up with schedule or budget goals was beneficial range, there is a range of limited difference in the
overwhelmed by the accumulated schedule and cost effect of pressure as the cycle time and effort curves flatten.
overruns, which further discouraged the team’s effort to The detrimental effects of extremely high levels of pressure
make adjustment. Meanwhile, adding personnel late in the are evident when pressure continues to increase beyond
project simply increased the cost overrun and further this range. To software management, the U-shaped view
delayed the schedule, consistent with Brook’s Law. and the beneficial or excessive range associated with it
The statistical analysis and the follow-up investigation provide valuable guidance for feasible budget and deadline
offered several important implications to research and setting policies. Our results suggest that, with a properly
practice. First, the finding about the development team’s tightened budget, software management can save time and
reaction to extremely low/high pressure implies the manpower simultaneously. The reduced cycle time under
generalizability of the theoretical development of this study budget pressure is not offset by an increased workforce.
(i.e., the Yerkes-Dodson Law and its underlying mechan- Management can use historical data to estimate the
isms) to the software development context. For example, the beneficial range of budget pressure. With the knowledge
development team’s reaction to low levels of pressure of beneficial pressure range, software management can be
conforms to the goal setting theory [44] that, for acceptable more effective in contract negotiations.
goals, goal difficulty has a positive effect on performance. The lack of support for the hypotheses regarding the
Meanwhile, high levels of pressure generate a negative effect of schedule pressure reveals a unique requirement for
reaction because they create too much challenge for the software projects to harvest positive effects of schedule
development team [14]. The finding about the development pressure: the cooperation between clients and software
team’s reaction also helps us gain deeper understanding of development teams. Observation from our research site
previous studies on schedule pressure. For example, indicates that clients may not sympathize with develop-
Putnam’s Rayleigh curve model [17] might have captured ment teams in schedule pressure. The behavior of clients
the negative reaction of management to high levels of can diminish the effect of schedule pressure found in
schedule pressure. At the same time, the seemingly contra- previous studies. For future research about software
dictory findings from Jeffery’s study [16] may be attributed development schedule, this study implies the importance
to the development team’s positive reaction to low levels of of examining client involvement as a significant factor in
schedule pressure. To practitioners, knowing that different development cycle time. Meanwhile, for software manage-
levels of budget or schedule pressure can generate dis- ment, it is critical to motivate the clients to share schedule
tinctive reactions can help them anticipate and interpret pressure and cooperate with development teams in redu-
development team performance. cing development cycle time.
Second, the strong statistical support for the U-shaped When interpreting the results, we were aware of several
relationships between budget pressure and development caveats. First, the results are based on data from a single
cycle time and effort suggests that in a software development organization. This could affect the generalizability of the
context, budget pressure may be a more robust indicator of findings, particularly the beneficial/excessive budget or
the impact of job-related pressure on performance. Schedule schedule ranges. However, pooling data from multiple
pressure did not exhibit similar effect because its effect on organizations could introduce confounding effects of
cycle time was reduced by the inelasticity of customer different estimation methods and different measures of
review activities; the effect on manpower was offset by key variables. Moreover, structural and environmental

Authorized licensed use limited to: Sreenidhi Institute of Science & Technology. Downloaded on November 5, 2009 at 04:49 from IEEE Xplore. Restrictions apply.
636 IEEE TRANSACTIONS ON SOFTWARE ENGINEERING, VOL. 35, NO. 5, SEPTEMBER/OCTOBER 2009

variations of different organizations could confound the REFERENCES


relationship between pressure and software development [1] D. Burdick, “Celestica Transforms Competitiveness with
performance. Future research that pools data from multiple C-Commerce,” Gartner Case Study, Jan. 2000.
organizations can seek to validate and extend the results of [2] D. Chatterjee, R. Grewal, and V. Sambamurthy, “Shaping Up for
this study by adequately controlling the confounding E-Commerce: Institutional Enablers of the Organizational Assim-
ilation of Web Technologies,” MIS Quarterly, vol. 26, no. 2, pp. 65-
variables discussed earlier. Second, the pressure measures 90, 2002.
only capture the postestimation budget and schedule [3] T.L. Friedman, The World Is Flat: A Brief History of the Twenty-First
compression rather than pressure arising during a project. Century. Farrar, Straus and Giroux, 2006.
That prevented us from examining the dynamic impacts of [4] B. Fitzgerald, “The Transformation of Open Source Software,”
MIS Quarterly, vol. 30, no. 3, pp. 587-598, 2006.
schedule and budget pressure throughout a project. In [5] B. Ives and S.L. Jarvenpaa, “Applications of Global Information
future work, budget and schedule pressure can be Technology: Key Issues for Management,” MIS Quarterly, vol. 15,
evaluated at several points during a development project. no. 1, pp. 33-50, 1991.
An interesting question to examine is whether the effect of [6] N. Levina and J.W. Ross, “From the Vendor’s Perspective:
Exploring the Value Proposition in Information Technology
pressure found in this study can be applied to later phases Outsourcing,” MIS Quarterly, vol. 27, no. 3, pp. 331-364, 2003.
of a project. [7] K.J. Stewart and S. Gosain, “The Impact of Ideology on
Effectiveness in Open Source Software Development Teams,”
MIS Quarterly, vol. 30, no. 2, pp. 291-314, 2006.
7 CONCLUSION [8] E. Yourdon, Death March: The Complete Software Developer’s Guide to
Surviving “Mission Impossible” Projects. Prentice Hall PTR, 1997.
In this paper, we formalize the nonlinear effect of manage- [9] R.M. Yerkes and J.D. Dodson, “The Relation of Strength of
ment pressures on project performance as U-shaped Stimulus to Rapidity of Habit Formation,” J. Comparative and
Neurological Psychology, vol. 18, pp. 459-482, 1908.
relationships. The relationships are tested with data from
[10] W.E. Scott, “Activation Theory and Task Design,” Organizational
a research site, where estimation bias is consciously Behavior and Human Performance, vol. 1, pp. 3-30, 1966.
minimized and information about the potential confound- [11] R.A. Dienstbier, “Arousal and Physiological Toughness: Implica-
ing effects such as software process, complexity, and tions for Mental and Physical Health,” Psychological Rev., vol. 96,
no. 1, pp. 84-100, 1989.
conformance quality is properly tracked. The data analysis [12] H. Anisman and Y. LaPierre, “Neurochemical Aspects of Stress
supports the U-shaped relationships between budget and Depression: Formulations and Caveats,” Psychological Stress
pressure and software development cycle time and devel- and Psychology, R.W. Neufeld, ed., McGraw-Hill, 1982.
[13] J. Schaubroeck and D.C. Ganster, “Chronic Demands and
opment effort. On the other hand, hypotheses about the Responsivity to Challenge,” J. Applied Psychology, vol. 78, no. 1,
U-shaped impacts of schedule pressure on development pp. 73-85, 1993.
cycle time and effort are rejected. [14] M. Frankenhaeuser and B. Gardell, “Underload and Overload in
Working Life: Outline of a Multidisciplinary Approach,” J. Human
In the highly competitive software engineering industry, Stress, vol. 2, no. 3, pp. 35-46, 1976.
companies are looking for ways to make software produc- [15] R.W. Zmud, “Management of Large Software Development
tion more efficient. The main contribution of this study is to Efforts,” MIS Quarterly, vol. 4, no. 2, pp. 45-55, 1980.
[16] D.R. Jeffery, “Time-Sensitive Cost Models in the Commercial MIS
rigorously assess the possibility of achieving improved
Environment,” IEEE Trans. Software Eng., vol. 13, no. 7, pp. 852-
cycle time and effort with proper management of pressure 859, July 1987.
in software development. The analysis suggests that there [17] L.H. Putnam, “A General Empirical Solution to the Macro
exists a beneficial range of budget pressure. Such beneficial Software Sizing and Estimating Problem,” IEEE Trans. Software
Eng., vol. 4, no. 4, pp. 345-360, July 1978.
effects may also be realized from schedule pressure if [18] B.W. Boehm, Software Engineering Economics. Prentice-Hall, 1981.
software clients cooperate with development teams in [19] F.P. Brooks, The Mythical Man Month: Essays on Software Engineer-
dealing with schedule compression. In addition to the ing. Addison-Wesley, 1974.
[20] C.S. Raghavendra and S. Hariri, “Reliability Optimization in the
empirical findings, this study sheds light on the funda- Design of Distributed Systems,” IEEE Trans. Software Eng., vol. 11,
mental mechanisms underlying the U-shaped relationships no. 10, pp. 1184-1193, Oct. 1985.
between development pressure and software performance. [21] G. Ruhe and M.O. Saliu, “The Art and Science of Software Release
Planning,” IEEE Software, vol. 22, no. 6, pp. 47-53, Nov./Dec. 2005.
By connecting software development with organizational [22] Q. Hu, R.T. Plant, and D.B. Hertz, “Software Cost Estimation
studies, this study not only helps to improve the general- Using Economic Production Models,” J. Management Information
izability of related findings from the software literature Systems, vol. 15, no. 1, pp. 143-163, 1998.
[23] B.A. Kitchenham, “Empirical Studies of Assumptions That
(e.g., [16], [17], [19], [23]) but also enriches the organiza- Underlie Software Cost-Estimation Models,” Information and
tional literature by extending its theories to the software Software Technology, vol. 34, no. 4, pp. 211-219, 1992.
context. For software management, the theoretical under- [24] R.D. Austin, “The Effects of Time Pressure on Quality in Software
standing as well as the empirical findings from this study Development: An Agency Model,” Information Systems Research,
vol. 12, no. 2, pp. 195-207, 2001.
can help them attain more effective negotiation strategies, [25] K.T. Abdel-Hamid and E.S. Madnick, “Lessons Learned from
deadline, and budget setting policies, which ultimately Modeling the Dynamics of Software Development,” Comm. ACM,
adds to the competitive advantage of a software company. vol. 32, no. 12, pp. 1426-1455, 1989.
[26] W.S. Humphrey, A Discipline for Software Engineering. Addison-
Wesley, 1995.
[27] R.D. Banker, G.B. Davis, and S.A. Slaughter, “Software Develop-
ACKNOWLEDGMENTS ment Practices, Software Complexity, and Software Maintenance
The authors would like to thank Dr. Dieter Rombach and Performance: A Field Study,” Management Science, vol. 44, no. 4,
pp. 433-451, 1998.
the three anonymous reviewers for their insightful com-
[28] K.T. Abdel-Hamid, “The Economics of Software Quality Assur-
ments. They would also like to thank Tara Thomas for her ance: A Simulation-Based Case Study,” MIS Quarterly, vol. 12,
assistance in data collection. no. 3, pp. 395-411, 1988.

Authorized licensed use limited to: Sreenidhi Institute of Science & Technology. Downloaded on November 5, 2009 at 04:49 from IEEE Xplore. Restrictions apply.
NAN AND HARTER: IMPACT OF BUDGET AND SCHEDULE PRESSURE ON SOFTWARE DEVELOPMENT CYCLE TIME AND EFFORT 637

[29] K.T. Abdel-Hamid, “The Dynamics of Software Projects Staffing: [58] L.G. Robert, “Evolving a New Theory of Project Success,” Comm.
A System Dynamics Based Simulation Approach,” IEEE Trans. ACM, vol. 42, no. 11, pp. 17-19, 1999.
Software Eng., vol. 15, no. 2, pp. 109-119, Feb. 1989. [59] N.E. Fenton and S.L. Pfleeger, Software Metrics: A Rigorous and
[30] K.T. Abdel-Hamid, K. Sengupta, and D. Ronan, “Software Project Practical Approach. Int’l Thompson Computer Press, 1997.
Control: An Experimental Investigation of Judgment with Fallible [60] R.D. Banker, H. Chang, and C. Kemerer, “Evidence on Economies
Information,” IEEE Trans. Software Eng., vol. 19, no. 6, pp. 603-612, of Scale in Software Development,” Information Software Technol-
June 1993. ogy, vol. 36, no. 5, pp. 275-282, 1994.
[31] K.T. Abdel-Hamid, K. Sengupta, and C. Swett, “The Impact of [61] R.D. Banker and S.A. Slaughter, “A Field Study of Scale
Goals on Software Project Management: An Experimental Economies in Software Maintenance,” Management Science,
Investigation,” MIS Quarterly, vol. 23, no. 4, pp. 531-555, 1999. vol. 43, no. 12, pp. 1709-1725, 1997.
[32] M.E. Helander, Z. Ming, and N. Ohlsson, “Planning Models for [62] A.J. Albrecht and J.E. Gaffney, “Software Function, Source Lines
Software Reliability and Cost,” IEEE Trans. Software Eng., vol. 24, of Code, and Development Effort Prediction: A Software Science
no. 6, pp. 420-434, June 1998. Validation,” IEEE Trans. Software Eng., vol. 9, no. 6, pp. 639-648,
[33] S. McConnell, “Feasibility Studies,” IEEE Software, vol. 15, no. 3, Nov. 1983.
pp. 119-120, May/June 1998. [63] C.F. Kemerer, Software Project Management Readings and Cases.
[34] K. Srinivasan and D. Fisher, “Machine Learning Approaches to McGraw-Hill, 1997.
Estimating Software Development Effort,” IEEE Trans. Software [64] M.S. Krishnan, C.H. Kriebel, S. Kekre, and T. Mukhopadhyay,
Eng., vol. 21, no. 2, pp. 126-137, Feb. 1995. “An Empirical Analysis of Productivity and Quality in Software
[35] J. Singh, “Striking a Balance in Boundary-Spanning Position: An Products,” Management Science, vol. 46, no. 6, pp. 745-759, 2000.
Investigation of Some Unconventional Influences of Role Stressors [65] D.E. Harter, M.S. Krishnan, and S.A. Slaughter, “Effects of Process
and Job Characteristics on Job Outcomes of Salespeople,” Maturity on Quality, Cycle Time, and Effort in Software Product
J. Marketing, vol. 62, no. 3, pp. 69-86, 1998. Development,” Management Science, vol. 46, no. 4, pp. 451-466,
[36] M.G. Weinberg, The Psychology of Computer Programming. Van Apr. 2000.
Nostrand Reinhold Company, 1971. [66] C. Jones, Applied Software Measurement: Assessing Productivity and
[37] H. Selye, The Stress of Life. McGraw-Hill, 1956. Quality. McGraw-Hill, 1996.
[38] R.G. Stennett, “The Relationship of Performance Level to Level of [67] M. Agrawal and K. Chari, “Software Effort, Quality, and Cycle
Arousal,” J. Experimental Psychology, vol. 54, no. 1, pp. 54-62, 1957. Time: A Study of CMM Level 5 Projects,” IEEE Trans. Software
[39] R.N. Berry, “Skin Conductance Levels and Verbal Recall,” Eng., vol. 33, no. 3, pp. 145-156, Mar. 2007.
J. Experimental Psychology, vol. 63, pp. 275-286, 1962. [68] M.C. Paulk, C.V. Weber, B. Curtis, and M.B. Chrissis, The
[40] E. Duffy, Activation and Behavior. Wiley, 1962. Capability Maturity Model: Guidelines for Improving the Software
[41] J.E. McGrath, “Stress and Behavior in Organization,” Handbook of Process. Addison-Wesley, 1995.
Industrial and Organizational Psychology, M.D. Dunnette, ed., [69] G. Peeters and G. Dewey, “Reducing Bias in Software Project
pp. 1351-1395, John Wiley, 1976. Estimates,” CrossTalk—The J. Defense Software Eng., pp. 20-24, Apr.
[42] H. Sjoberg, “Interaction of Task Difficulty, Activation and Work 2000.
Load,” J. Human Stress, vol. 3, no. 1, pp. 33-38, 1977. [70] R.D. Banker and C.F. Kemerer, “Scale Economies in New Software
Development,” IEEE Trans. Software Eng., vol. 15, no. 10, pp. 416-
[43] J.R.P. French Jr., R.D. Caplan, and W. Harrison, The Mechanisms of
429, Oct. 1989.
Job Stress and Strain. John Wiley, 1982.
[71] R. Davidson and J. MacKinnon, “Several Tests for Model
[44] M. Erez and I. Zidon, “Effects of Goal Acceptance on the
Specifications in the Presence of Multiple Alternatives,” Econome-
Relationship of Goal Setting and Task Performance,” J. Applied
trica, vol. 49, no. 3, pp. 781-793, 1995.
Psychology, vol. 69, no. 1, pp. 69-78, 1984.
[72] R.D. Cook and S. Weisberg, Applied Regression Including Computing
[45] M. Csikszentmihalyi, “Play and Intrinsic Rewards,” J. Humanistic
and Graphics. John Wiley and Sons, 1999.
Psychology, vol. 15, no. 3, pp. 41-63, 1975.
[73] N. Smirnov, “On the Estimation of the Discrepancy between
[46] M. Csikszentmihalyi, Beyond Boredom and Anxiety. Jossey-Bass, Empirical Curves of Distribution for Two Independent Samples,”
1977. Bull. Math. Univ. of Moscow, vol. 2, no. 1, pp. 3-16, 1939.
[47] E.A. Locke and G.P. Latham, “Building a Practically Useful [74] H. White, “Heteroskedasticity-Consistent Covariance Matrix
Theory of Goal Setting and Task Motivation,” Am. Psychologist, Estimator and a Direct Test for Heteroskedasticity,” Econometrica,
vol. 57, no. 9, pp. 705-717, 2004. vol. 48, no. 5, pp. 817-838, 1980.
[48] C.N. Parkinson, Parkinson’s Law and Other Studies in Administra-
tion. Random House, 1957.
Ning Nan received the PhD degree in business
[49] G. Gutierrez and P. Kouvelis, “Parkinson’s Law and Its Implica-
information technology from the Ross School of
tions for Project Management,” Management Science, vol. 37, no. 8,
Business at the University of Michigan in 2006.
pp. 990-1001, Aug. 1991.
She is currently an assistant professor in the
[50] R. Rodgers and J. Hunter, “Impact of Management by Objectives Price College of Business at the University of
on Organizational Productivity,” J. Applied Psychology, vol. 76, Oklahoma. Her research interests include man-
no. 2, pp. 322-336, 1991. agement of software development projects, in-
[51] A. O’Leary-Kelly, J. Martocchio, and D. Frink, “A Review of the group behaviors and time zone effect on global
Influence of Group Goals on Group Performance,” Academy of software development teams, and postadoptive
Management J., vol. 37, no. 5, pp. 1285-1301, 1994. information technology use behaviors.
[52] R. Baum, E. Locke, and K. Smith, “A Multi-Dimensional Model of
Venture Growth,” Academy of Management J., vol. 44, no. 2, pp. 292-
303, 2001.
[53] K.H. Price, “Decision Responsibility, Task Responsibility, Iden- Donald E. Harter received the MSc and PhD
tifiability, and Social Loafing,” Organizational Behavior and Human degrees in management information systems
Decision Processes, vol. 40, no. 3, pp. 330-345, 1987. from Carnegie Mellon University in 1996 and
[54] S. Ang and S.A. Slaughter, “Work Outcomes and Job Design for 2000, respectively. He is currently an assistant
Contract versus Permanent Information Systems Professionals on professor in the Whitman School of Manage-
Software Development Teams,” MIS Quarterly, vol. 25, no. 3, ment at Syracuse University. His research
pp. 321-350, 2001. interests include measurement of the effect
[55] S. Nollen and H. Axel, Managing Contingent Workers. Am. of software processes on software develop-
Management Assoc., 1996. ment costs, schedules, quality, support activ-
[56] J. Pfeffer and J. Baron, “Taking the Workers Back Out: Recent ities, and software personnel compensation.
Trends in the Structuring of Employment,” Research in Organiza- He is a member of the IEEE.
tional Behavior, B. Staw and L.L. Cummings, eds., JAI Press, 1988.
[57] D.P. Darcy, C.F. Kemerer, S.A. Slaughter, and J.E. Tomayko, “The
Structural Complexity of Software: Testing the Interaction of . For more information on this or any other computing topic,
Coupling and Cohesion,” IEEE Trans. Software Eng., vol. 31, no. 11, please visit our Digital Library at www.computer.org/publications/dlib.
pp. 982-995, Nov. 2005.

Authorized licensed use limited to: Sreenidhi Institute of Science & Technology. Downloaded on November 5, 2009 at 04:49 from IEEE Xplore. Restrictions apply.

Anda mungkin juga menyukai