Discussion 2
Performance Framework
A. Program Logic
What are the aspects of program or organizational performance that would be
important to monitor on a regular basis in a performance management system? What does
the effectiveness of a particular public program or the quality of the services it provides
entail? What does efficiency mean with respect to a given program, or the productivity of
employees working in a service delivery system? In order for a measurement system to
be useful, it must focus on the most appropriate aspects of performance. To ensure that a
measurement system is results oriented as opposed to being data driven, it should be
guided by a clear performance framework, a structure that identifies the parameters of
program or organizational performance and may outline the presumed relationships
among them. This chapter addresses the what of performance measurement: What are the
important dimensions of program performance to address, and what are the principal
kinds of performance measures to track them? More specifically, the chapter discusses
one principal type of performance framework, program logic models. These logic models
focus primarily on program performance, reflect a results-oriented perspective, and are
appropriate for programs delivered by both public and nonprofit organizations. Chapter 4
then discusses goal structures as performance frameworks, a very different yet
complementary, and by no means incompatible, type of approach.
Developing useful measures of program performance requires a clear
understanding of what a program does and the results it is intended to accomplish
(Poister, 1978; Wholey, 1979; Broom, Harris, Jackson, & Marshall, 1998; Sowa, Selden,
& Sandfort, 2004; McDavid & Hawthorn, 2006). Program logic models are schematic
diagrams that represent the logic underlying a program’s design, indicating how various
components are expected to interact, the goods or services they produce, and how they
generate the desired results—the logic by which program activities are expected to lead
to targeted outcomes (Poister, 1978; Poister, McDavid & Magoun, 1979; Hatry, Van
Houten, Plantz, & Greenway, 1996; Funnell & Rogers, 2011; Knowlton & Phillips,
2012). Clarifying desired outcomes and the underlying logic by which they are expected
to be achieved is essential for effective performance management, and a good logic
model can provide the scaffolding for building a performance measurement system
(Frechtling, 2007). Once a logic model has been developed and adopted, the relevant
performance measures can be identified systematically and confidently.
Public programs should be planned and managed with an eye toward specifying
and achieving desirable results. They should be viewed as interventions involving service
delivery or enforcement activity designed to address some problem, meet some need, or
have a favorable impact on some unsatisfactory condition in a way that has been defined
as serving the public interest. The positive impacts so generated constitute the program’s
intended results, which would justify support for the program in the first place. A
program’s intended results, or its outcomes, occur “out there” in the community, within a
targeted area or target population, or across the nation or state or local jurisdiction
generally, not inside the program itself or the agency or organizational unit that operates
it. Obviously the intended results should be clearly understood and monitored on a
regular basis. If a programmatic entity cannot articulate worthwhile results and provide
evidence that programmatic activity is indeed producing them, continued support should
be questioned at the very least.
Resources are used to carry on program activities and provide services that
produce immediate products, or outputs. These outputs are intended to lead to outcomes,
which are the substantive changes, improvements, or benefits that are supposed to result
from the program. Frequently these outcomes themselves occur in sequence, running
from initial outcomes to intermediate and longer-term outcomes. Usually the logic
underlying a program design is also predicated on a flow of customers who are served by
a program or a set of cases the program deals with. In addition, it is important to
recognize the external factors in a program’s environment or operating context, which
may influence its performance.
The sets of activities that make up the work of most public programs involve the
provision of services or the enforcement of laws or regulations (or both). For example,
the principal activities in a neighborhood health clinic might include conducting physical
examinations and well-baby checks, giving inoculations, and prescribing treatments and
medications for illnesses and chronic conditions; in the criminal investigations unit of a
local police department, the principal activities would include examining crime scenes,
interviewing witnesses, examining physical evidence, checking out leads, and gathering
additional information. These programmatic activities and the outputs they produce need
to be identified clearly, whether they are carried on by public sector employees working
in the program or by private firms or nonprofit organizations that are contracted to carry
out service delivery. The principal resources that most public and nonprofit programs use
are personnel, physical facilities, equipment, materials, and contract services. Personnel
may include volunteers as well as employees, and sometimes it is helpful to break them
down into occupational categories, tracking, for example, the numbers of uniformed
patrol officers, detectives, crime lab personnel, and support staff in order to gauge labor
productivity in a local police department.
In many public programs, especially those carried on by productiontype agencies,
the work performed and the results obtained apply to cases or groups of cases that come
into the program or are treated by the program in some fashion. Frequently the cases are
the program’s primary customers (or consumers or clients). This is almost always true for
human service and educational programs—patients treated in public hospitals, children
served by a foster care program, clients aided in a counseling program, or students
enrolled in a community college, for example—but customers are also often the principal
cases in other types of programs, for instance, disabled persons using demand-responsive
transportation services or the number of families living in dwelling units provided by a
public housing authority. However, with some programs, the most likely definition of
cases may be something other than customers. For example, while the customers of a
state highway maintenance program are individual motorists, it makes more sense to
think of the cases to be processed as consisting of road miles or road segments to be
maintained. Similarly, the implicit customer of the Keep Nebraska Beautiful program is
the public at large, but the “cases” treated by this program consist of small, targeted
geographical areas. While the cases processed by a state’s driver’s license permitting
program are individual applicants, the customers, the cases processed by a state’s vehicle
registration program, are probably best defined as the vehicles rather than the customers
who are registering them. Often a public or nonprofit program may define its cases in
more than one way. For example, the US Internal Revenue Service may consider the
individual customer as the case with respect to its tax preparation assistance function but
focus on the tax return as the case in terms of its collections and auditing functions.
It is important to identify external influences in thinking about a program’s logic
because they may be critical in either facilitating or impeding success. Many of these
external influences concern client characteristics or the magnitude or severity of need for
the program, but they are by no means limited to that. Any factor or condition—physical,
social, economic, financial, psychological, or cultural—that is likely to influence program
performance and is largely beyond the control of the program or agency may be relevant
to track as an external influence. For example, winter weather conditions may explain
differences in the performance of a highway maintenance program from year to year,
while differences in labor market conditions may explain differences in the effectiveness
of similar job training programs in different localities, and variation in local industrial
base, land use patterns, and commuting behavior are likely to influence the federal
Environmental Protection Agency’s success in enforcing clean air standards in different
parts of the country. Such external factors are important to take into account in clarifying
a program’s underlying logic because they can be extremely helpful in interpreting the
meaning of performance data.
The most important distinction to be made in identifying program logic is that
between outputs and outcomes. Outputs represent what a program actually does and what
it produces directly; outcomes are the results it generates. Operations managers
appropriately focus on the production of high-quality outputs in an efficient manner, but
managers who are concerned with overall performance must look beyond outputs to
outcomes because they represent program effectiveness. In terms of program logic,
outputs have little inherent value because they do not constitute direct benefits, but they
are essential because they lead directly to these benefits or trigger the causal sequences of
changes that lead to the desired results. Outputs are best thought of as necessary but
insufficient conditions for success. They are the immediate products or services produced
by a program, and without an appropriate mix and quality of outputs, a program will not
be able to generate its intended results. However, if the underlying program logic is
flawed—if the assumptions of causal connections between outputs and results do not hold
up in reality—the desired outcomes will not materialize, at least not as a result of the
program. Usually the production of outputs is largely, although not exclusively, under the
control of program managers, but outcomes tend to be influenced more strongly by a
wider array of external factors beyond the program’s control. Thus, the production of
outputs is no guarantee that outcomes will result, and it is important therefore to measure
outcomes directly in order to monitor program performance.
Outputs often represent the amount of work performed or the volume of activity
completed, such as hours patrolled by the police, miles of highway constructed, AIDS
education seminars conducted or antibody tests given, or the number of vocational
training classes conducted in a juvenile justice boot camp. Sometimes outputs are
measured in terms of the number of clients or cases treated, for example, the number of
crimes investigated or calls for service responded to by the police, the number of AIDS
patients given treatment or counseling, or the number of youths discharged from juvenile
boot camps. Outcomes are the substantive results generated by producing these outputs.
Criminal investigations and arrests do not really count for much— for instance, if the
police are not able to solve the crimes they are working on, and reconstructed highway
segments do not serve any particular public interest or create value if they do not result in
improved flow of traffic and reduced travel times for the motorists using them. Similarly,
AIDS-awareness seminars are not particularly worthwhile if they do not lead to decreases
in the kinds of risky behavior—unprotected sex and use of dirty drug needles, for
example—that spread the HIV virus. And training units and hours spent in after-care
activity are not effective in attaining their rehabilitative purposes if the youths discharged
from boot camps are not productively engaged in school or work and refraining from
further criminal activity. Outcomes are the ultimate criteria for gauging program
effectiveness, but as direct products of program activity, outputs are critical for achieving
intended outcomes.
The distinction between outputs and outcomes may also vary depending on the
perspective from which they are viewed. For a staff support function such as policy
research, planning, and evaluation in a federal agency, for example, typical outputs might
be policy analyses completed, program designs revised, needs assessments completed,
and program evaluations conducted. In some cases, producing these outputs might be
intended to lead to such outcomes as additional legislation passed or new programs
authorized by the Congress. Whereas the implementation of policy changes, revised
program strategies, or service delivery arrangements by the agency in response to
recommendations from the policy research would constitute outputs for the agency as a
whole, these same measures would constitute outcomes for the office of policy research,
planning, and evaluation itself.
B. Diverse Logic Models
Outputs, outcomes, and other elements can be identified in logic models that may
be as general or detailed, or as simple or complex, as needed. Although it is always a
mistake to bend reality to fit a preconceived model, the kind of program logic models
presented here are quite flexible and can be adjusted to represent any public or nonprofit
program. For example, the set of program components to be included can range from
only one to numerous activities, and the connections between outputs and desired
outcomes may be very direct, or they can occur through numerous initial and
intermediate results. Similarly, the strands of logic that connect outputs to various
outcomes can converge at different points and in different sequences along the way.
Although these models show the logic generally moving from left to right, they are not
necessarily designed to represent the chronological order in which treatments are
provided. These are logic models, not flowcharts that show the sequence in which
individual cases move through a system.
The crisis stabilization unit produces a variety of outputs that reflect the work
performed in providing services, such as medical assessments and nursing assessments
conducted, physical examinations conducted, medical detoxifications completed,
psychiatric assessments, education program modules, therapy sessions, Alcoholics
Anonymous (AA) meetings, and referrals or placements. The initial outcomes produced
by these service outputs are substance abuse consumers who have been stabilized through
detoxification without physiological withdrawal symptoms and the number of psychiatric
consumers who have been stabilized with medication. A complementary initial outcome
is consumers who have been empowered through counseling, educational programs, and
support groups to make more responsible decisions regarding their own behavior. For the
substance abusers, the intermediate outcomes are that they enter appropriate long-term or
day-patient treatment programs and continue to abstain from drugs and alcohol over the
long run. For the psychiatric consumers, the intermediate outcomes are that after being
discharged from the unit, they return to baseline or desired behavior levels and continue
to take their appropriate medications. For both clientele groups, the intended longer-term
outcomes are that they resume normal patterns of work, family, and community life and
that all of this results in reduced needs for acute care for them.
The initial outcome of all this activity is that clients have developed the
knowledge and skills needed to engage in occupations that are viable for them and that
they apply for suitable jobs in the competitive marketplace or, in some cases, sheltered
workshops. The intermediate outcome is that the clients are placed in suitable jobs. Once
clients have secured suitable jobs, the program may provide on-the-job evaluations with
recommendations to assist them in adjusting to new jobs. This is all aimed at helping
clients to continue working in suitable jobs and being successfully employed over the
long run. To the extent this longer-term outcome is achieved, the program’s mission is
being met effectively.
Individual public programs often cluster into larger sets of related programs in
which their results chains interact in pursuit of broader goals even though they are
delivered separately by different agencies or organizational units. For example, figure 3.4
shows a logic model for a state government’s highway safety program consisting of four
major components: driver licensing, highway patrol, safety promotion, and traffic
engineering. In some states, all four of these functions are the responsibility of a single
agency, but that is often not the case. In Georgia, for instance, driver licensing is the
responsibility of the Department of Driver Services, highway patrol is performed by the
Georgia State Patrol, safety promotion efforts are led by the Governor’s Office of
Highway Safety, and traffic engineering is within the purview of the Georgia Department
of Transportation.
However, since these four functions are collectively concerned with reducing
highway crashes, injuries, and fatalities resulting from crashes and the cost and suffering
that derives from them, a central executive agency such as a state’s office of planning and
budget might well benefit from monitoring their performance on a coordinated basis.
While each of these agencies delivers its own outputs, which are expected to generate
particular kinds of initial outcomes, the driver licensing, highway patrol, and safety
promotion programs are all aimed at ensuring safe driving behavior as an intermediate
outcome, and the safety promotion program is also expected to lead directly to fewer
highway-related injuries and fatalities, while the traffic engineering program is intended
to lead directly to the longer-term outcome of fewer highway crashes. The combined
logic also holds that the first three programs will contribute indirectly to fewer crashes,
and all four programs should lead to the longer-term outcome of reduced costs and
suffering resulting from highway crashes. Tracking the performance of these four
programs in concert as they form a larger system provides a more comprehensive picture
of the effectiveness of a state’s overall highway safety program.
Traditionally sexually transmitted disease (STD) prevention programs in the
United States operated through direct service delivery—screening, diagnosis, treatment,
and partner services—provided by local STD clinics supported in part with funding from
the US Centers for Disease Control (CDC) primarily through state health departments. A
decade or so ago, however, as a result of environmental and program assessments, CDC
staff became convinced that this approach was no longer adequate, due largely to
increasingly fragmented health care delivery systems, the lack of coordination among
related health programs at the local level, and the fact that the clinics did not have a
strong track record in reaching some of the most critical target populations. Thus, they
concluded that STD prevention programs needed to implement a broader range of
strategies to leverage impact on a variety of other stakeholders, such as managed care
organizations, private medical practices, schools, detention and corrections facilities, and
community-based organizations in order to reach out to at-risk persons and effectively
contain the spread of syphilis, gonorrhea, and chlamydia.
Shifting focus to earlier-stage outputs and immediate outcomes allows the Centers
for Disease Control and Prevention (CDC) to monitor how state and local STD
prevention programs are influencing the larger public health system. This strategic shift
involves tracking activities and outputs not only from local grantees directly involved in
STD prevention but also from key institutions such as schools, health networks, and even
jails. By monitoring outputs such as the number of educational sessions conducted in
schools, distribution of condoms in health networks, or STD screenings in jails, the CDC
gains insight into the reach and intensity of STD prevention efforts across various
settings. These outputs serve as early indicators of program implementation and activity
levels. Immediate outcomes, such as changes in knowledge, attitudes, and behaviors
related to STD prevention, provide early signals of the program’s effectiveness. For
example, increases in condom use among adolescents or improved adherence to STD
treatment protocols in correctional facilities indicate progress toward reducing
transmission rates and improving health outcomes.
Monitoring how STD prevention activities extend beyond grantee organizations
to influence broader systems—like educational institutions and healthcare settings—helps
assess the program’s systemic impact. This approach acknowledges the
interconnectedness of health promotion efforts and encourages collaborative efforts
across sectors to achieve public health goals. Requiring local grantees to report on their
own activities as well as activities within partner institutions fosters accountability and
transparency. It ensures that program resources are effectively utilized and encourages
grantees to leverage partnerships to maximize program reach and impact. Collecting data
on STD prevention activities and outcomes enables data-driven decision-making at both
local and national levels. Analysis of trends and outcomes informs strategic planning,
resource allocation, and policy development to strengthen STD prevention efforts and
address emerging challenges.
Engaging schools, health networks, and correctional facilities in STD prevention
activities builds their capacity to deliver effective interventions and integrate health
promotion into their respective settings. Collaboration enhances resource sharing,
expertise exchange, and coordinated responses to STD prevention and control. Ongoing
monitoring of outputs and outcomes supports continuous program improvement.
Identifying successful strategies and areas needing enhancement allows for timely
adjustments to intervention approaches, ensuring programs remain responsive to evolving
community needs and public health priorities. By emphasizing the monitoring of outputs
and immediate outcomes within state and local STD prevention programs, the CDC
enhances its ability to assess program impact, foster collaboration across sectors, and
promote effective public health interventions. This approach not only strengthens STD
prevention efforts but also contributes to broader health equity goals by addressing
disparities and promoting comprehensive health promotion strategies.
C. Performance Measures
The purpose of developing a logic model is to clarify what goes into a program,
identify its customers, pinpoint the services it provides, identify the immediate products
or outputs it produces, and specify the outcomes it is supposed to generate. Once this
logic has been articulated, in a narrative or in a schematic, or both, the most relevant
measures of program performance can be identified on a systematic basis. Although they
are often combined into different categories, for the most part the relevant types of
performance measures are measures of outputs, efficiency and productivity, service
quality, outcomes, cost-effectiveness, and customer satisfaction. One additional type of
performance measure, system productivity indicators, can also be included in many
measurement systems. Depending on the purpose of a given performance measurement
system and the level of detail on which the monitoring may focus, various of these types
of performance measures will be of paramount importance, but it usually makes sense to
consider all of them in designing a performance measurement system. For any given
program, all of these types of performance measures can generally be derived directly
from the logic model. In addition, other types of measures, in particular resources and
workload measures, are often monitored on a regular basis in performance measurement
systems, even though they are not usually considered to be performance measures in their
own right. Beyond performance measures, external variables representing environmental
or contextual factors, including needs indicators that are likely to influence programmatic
results, might also be included in performance measurement systems.
Output measures are important because they represent the direct products of
public agencies or programs. They often measure volumes of programmed activity, such
as the number of training programs conducted by a job training program, the number of
seminars presented by an AIDS prevention program, the miles of new four-lane highways
constructed by a state transportation department, or the hours of routine patrol logged in
by a local police department. Outputs are often measured in terms of the amount of work
that is performed—for example, the number of detoxification procedures completed by a
crisis stabilization unit, the number of job interviews arranged for clients of a vocational
rehabilitation program, or the gallons of patching material used on roads by highway
maintenance crews. Finally, output measures sometimes represent the number of cases
that are dealt with by a program, such as the number of flight segments handled by the
nation’s air traffic control program, the number of AIDS clients who receive counseling,
or the number of crimes investigated by the police.
Outputs are sometimes measured at different stages of the service delivery
process, and we can think of outputs chains occurring in some programs. For instance,
the outputs of crime investigation are usually specified as the number of initial responses
to crimes reported, the number of crimes investigated, and the number of arrests made.
Juvenile justice boot camps often measure the numbers of juveniles under their charge
who complete various training modules and receive other services and the number of
juveniles who are discharged from the camps, as well as the number of after-care visits or
activities reported. All of these stages of outputs are relevant to track because they
provide some indication of the amount of activity or work completed or the number of
cases being treated in some way.
Efficiency measures focus on the operating efficiency of a program or
organization, relating outputs to the resources used in producing them. They are most
frequently operationalized as unit cost measures expressed as the ratio of outputs to the
dollar cost of the collective resources consumed in producing them. Thus, the cost per
crime investigated, the cost per highway project design completed, the cost per AIDS
seminar conducted, the cost per ton of residential refuse collected, and the cost per
training program completed are all standard efficiency measures. While the operating
efficiency of the air traffic control program could be measured by the cost per flight
segment handled, a measure of the efficiency of a state board of nursing’s disciplinary
program would be the cost per investigation completed.
It may be appropriate to track a variety of efficiency measures for a given
program. For example, the cost per psychiatric assessment completed, the cost per
detoxification procedure conducted, the cost per therapy session conducted, and the cost
per support group meeting might all be relevant for a crisis stabilization unit if it has an
activitybased accounting system that can track the actual costs for these separate
activities. More general measures are often employed, such as the cost per highway lane
mile maintained or the cost per case in a child support enforcement program, but they are
really based more on workload than outputs. One particular efficiency measure that is
often used along these lines is the per diem—the cost per client per day in such
residential programs as hospitals, crisis stabilization units, juvenile detention centers, and
group homes for mentally disabled persons operated by nonprofit agencies.
While operating efficiency is most commonly expressed by these unit cost
measures, however, it is also measured in terms of cycle times, the average time required
to produce a single unit of output. With respect to a driver licensing program, for
example, the process for renewing licenses in a given drivers’ licensing center might
generate a renewed license every two minutes on average, while the average time
required for a state board of nursing to complete an investigation regarding a reported
violation of the authorized scope of practice for nurses licensed in a given state might be
168 days. Such time-based efficiency measures can also be expressed as the volume of
output produced in a certain period of time, for example, the feet of guardrail installed
per day by a highway maintenance program or the number of claims cleared per month
by a disability adjudication process. Productivity indicators are a special type of
efficiency measure that focus on the rate of output production per some specific unit of
resource, usually staff or employees. Since public service delivery tends to be labor
intensive, labor productivity measures are prevalent in performance monitoring systems
focusing on the production of outputs. To be meaningful, they also must be defined in
terms of some particular unit of time. For example, the number of flight segments
handled per air traffic controller per hour and the number of lane-miles of highway
resurfaced per maintenance crew per day are typical measures of labor productivity, as is
the number of nursing violation investigations completed per investigator per year.
In some cases, labor productivity ratios use the unit of measurement in both the
numerator and denominator, for example, the number of task hours completed per
production hour worked on a highway maintenance activity or the number of billable
hours of work completed per production hour worked in a state government printing
plant. Beyond labor productivity, in some cases the specific resource used as the basis for
a productivity indicator may measure equipment rather than personnel, for example, the
number of standard images printed per large press per hour in a government printing
office or the number of revenue vehicle miles operated per month per bus in a public
transit agency’s fleet. Staff-to-client ratios are sometimes loosely interpreted as
productivity measures, but this may be misleading. For example, the number of in-house
consumers per full-time staff member of a crisis stabilization unit may represent
productivity because those consumers are all receiving treatment. However, the number
of cases per adjuster in a state workers’ compensation program does not really provide
much information about the productivity of those employees because some or many of
those clients or cases may generate very little, if any, activity. The number of clients per
employee in a vocational rehabilitation program may not be particularly useful either,
again because the services being provided vary so widely from one client to the next, but
the number of clients counseled per vocational rehabilitation counselor would be more
meaningful because it represents the amount of work performed per staff member.
The concept of quality pertains most directly to service delivery processes and
outputs because they define the service that is being provided. When we think about
measuring outputs, we tend to think first of quantity—how much service is being
provided—but it is equally important to examine the quality of outputs as well. However,
this is not primarily a distinction between hard and soft measures. While service quality
is usually assessed subjectively at an individual level, performance measurement systems
track quality using more objective, quantitative data in the aggregate. The most common
dimensions of the quality of public and nonprofit services are turnaround time, accuracy,
thoroughness, accessibility, convenience, courtesy, and safety. For example, people who
are trying to renew their driver’s license tend to be most concerned about the accessibility
of the location where they do this, the convenience afforded in completing the process,
the total time including waiting time that it takes to complete the transaction, and, of
course, the accuracy of the paperwork that is processed (so that they won’t have to return
or repeat part of the process). In the Federal Aviation Administration’s air traffic control
program, the most important indicator of service quality is the number of controller errors
(instances in which controllers allow pilots to breach minimum distances to be
maintained between airplanes) per 1 million flight segments handled.
Frequently measures of service quality are based on standard operating
procedures that are prescribed for service delivery processes. Quality ratings of highway
maintenance crews, for instance, are usually defined by the extent to which the
establishment of the work site, handling of traffic through or around the work site, and
the actual work of patching potholes or resurfacing pavement comply with prescribed
operating procedures for such jobs. Juvenile justice detention centers have operating
procedures regarding such processes as safety inspections, fire prevention, key control,
perimeter checks, the security of eating utensils, supervision, and the progressive use of
physical force or chemical agents in order to ensure the security of the facility and the
safety of the juveniles in their custody. Quality assurance ratings are really compliance
measures, defined as the extent to which such processes are performed in compliance
with prescribed procedures. Yet other quality indicators, such as the number of escapes
from juvenile detention facilities or reported instances of child abuse, probably more
meaningful in terms of overall program performance, are defined more directly in terms
of desired outputs, juveniles detained safely and securely in this example.
It is fair to say that outcome measures constitute the most important category of
performance measures because they represent the degree to which a program is producing
its intended outcomes and achieving the desired results. These may relate to initial,
intermediate, or longer-term outcomes. Outcome measures for the air traffic control
program, for example, might include the number of near misses reported by pilots, the
number of midair collisions, and the number of fatalities per 100 million revenue
passenger–miles flown. The most important outcome measures tie back to the basic
purpose of a program. For example, the crisis stabilization unit exists to stabilize persons
with psychiatric or drug-induced mental crises and help them modify behaviors in order
to avoid falling into these same circumstances again. Thus, a key effectiveness measure
might be the percentage of all initial admissions that constitute readmissions within thirty
days. Similarly, the most important indicator of the effectiveness of a vocational
rehabilitation program is probably the number or percentage of clients who have been
successfully employed in the same job for six months. Along these same lines, the most
relevant effectiveness measures for a juvenile detention center are probably the
percentage of discharged youth who are attending school or engaged in gainful
employment and the percentage who have not recidivated back into the criminal justice
system within one year of having been discharged. Effectiveness measures for an AIDS
prevention program would be likely to include morbidity and mortality rates for AIDS,
along with the percentage of newborn babies who test positive for HIV.
Whereas indicators of operating efficiency are unit costs of producing outputs,
cost-effectiveness measures relate cost to outcome measures. Thus, for the crisis
stabilization unit, cost-effectiveness would be measured as the cost per stabilized
consumer. For the vocational rehabilitation program, the most relevant indicators of cost-
effectiveness would be the cost per client placed in suitable employment and the cost per
client successfully employed for six months or more. The cost-effectiveness of criminal
investigation activity would probably be measured as the cost per crime solved.
Effectiveness measures often become more esoteric and present more difficult
methodological challenges in operationalizing indicators. For example, the cost-
effectiveness of highway construction might well be conceptualized as the cost per
person-hour of reduced travel time, while the most relevant cost-effectiveness indicator
for an AIDS prevention program would probably be the cost per AIDS fatality avoided.
Both of these make complete sense in terms of program logic, but they are difficult to
operationalize.
Measures of customer satisfaction are often closely related to service quality
indicators, but the two are not identical and should be considered separate categories of
performance measures. Similarly, customer satisfaction measures are often associated
with effectiveness measures, but they provide a different perspective on overall program
performance. For example, measures of customer satisfaction with a vocational
rehabilitation program might be based on data from client evaluation forms asking how
satisfied they were with training programs they participated in, counseling services they
received, and assistance that was provided to them in finding a job. These all focus on
program outputs. In addition, clients who have been placed in jobs might be surveyed
after several months to assess their satisfaction with these jobs, focusing on program
effectiveness. These customer satisfaction ratings may or may not square with more
tangible measures of program outputs and effectiveness, but they do provide a
complementary perspective.
One way of gauging customer satisfaction is to track complaints. For example, a
public library system might monitor the number of complaints received from patrons per
week in each branch library. Second, some public and nonprofit agencies use customer
response cards to solicit immediate feedback regarding specific instances of service
delivery. A government printing office, for instance, might track the percentage of its
customers who rate their products as “good” or “excellent.” Probably the most frequently
used means of soliciting customer feedback is the customer survey, for example, the
percentage of victims reporting that they were “satisfied” with the initial police response
to their case. Similarly, a crisis stabilization unit might track the percentage of consumers
rating their services as “good” or “excellent,” while a highway maintenance operation
might estimate the percentage of motorists who are “satisfied” or “very satisfied” with
the condition of the roads they travel on.
Although the term does not appear often in typologies of performance measures,
another useful type of performance measure consists of system productivity measures.
These measures examine the ratio of outcome measures to related output measures in
order to gauge the effectiveness of a public agency or program in converting outputs to
outcomes. For example, public transit agencies often monitor the number of passenger
trips carried on the system, the principal outcome measure, per vehicle mile or vehicle
hours operated, which are standard output measures. Along the same lines, police
agencies can track the number of convictions obtained (outcome) to the number of arrests
made or cases solved (output), and child support enforcement programs can measure the
number of cases with support payment obligated per absentee parent located, while
housing rehabilitation programs can monitor the number of dwelling units brought into
compliance with codes per loan made under the program. In each of these cases, the
system productivity measure provides a clear indication of the extent to which producing
specified outputs leads to the accomplishment of desired outcomes.
Two other types of indicators, resource and workload measures, are usually not
thought of as performance measures in their own right, but they are often used in
computing other performance measures and are sometimes used in conjunction with other
performance measures. All the various types of resources supporting a program can be
measured in their own natural measurement units—for example, number of teachers,
number of school buildings or classrooms, number of computer work stations in a local
school system—or they can be measured and aggregated in their common measurement
unit, which is dollar cost. Although resource measures constitute investment at the front
end rather than something produced by the program, when managerial objectives focus
on improving the mix or quality of resources—maintaining a full complement of
teachers, for instance, or increasing the percentage of teachers with a master’s degree—
then it may be appropriate to track resource measures as indicators of performance.
However, the principal use of resource measures in tracking program performance is as a
basis for computing efficiency measures, such as the cost per hour of classroom
instruction, or costeffectiveness measures, such as the cost per student graduated.
Workload measures are often of great concern to managers because they represent
the flow of cases into a system or numbers of customers who need to be served. When
work standards are in place or average productivity rates have been established, workload
measures can be defined to represent resource requirements or the backlog of work in a
production system—for example, the number of production hours needed to complete all
jobs in the queue in a government printing office or the number of crew-days required to
complete all the resurfacing projects that would be needed to bring a city’s streets up to
serviceable standards. In some cases, when managerial objectives focus on keeping
workloads within reasonable limits—not exceeding two workdays pending in a central
office supply operation, for example, or keeping the workweeks pending within two
weeks in a disability determination program, or reducing the number of cases pending in
a large county’s risk management program by closing more cases than are opened in each
of the next six months— then workload measures may appropriately be viewed as
performance measures.
In addition to actual performance measures, it is often helpful for performance
monitoring systems to track other measures—variables external to a program itself—
which are likely to influence programmatic outcomes and perhaps the production of
outputs, efficiency, and service quality as well. For example, the principal outcomes of
public transit systems have to do with ridership measured by such indicators as passenger
trips carried per month, and ridership is expected to be influenced by such programmatic
factors as the amount of transit service provided, the quality of that service, and the fares
that are charged. However, other environmental factors, such as unemployment rates,
automobile ownership, and the price of gasoline, are likely to exert strong influence on
the ability of a transit system to attract ridership as well. Similarly, the success of a city’s
housing rehabilitation program in raising the percentage of dwelling units that meet
decent, safe, and sanitary conditions may depend heavily on the quality of the existing
housing stock, the percentage of absentee ownership, socioeconomic characteristics,
percentage of vacant properties, the quality of neighborhood schools, and the extent of
community cohesion in targeted areas.
However, such external factors that are well beyond the control of the program
may have an overwhelming influence on its ability to generate the desired kinds of
outcomes. Thus, although these external factors do not constitute part of the program
logic itself, logic models should not be developed in ignorance of these factors, and it can
be very helpful to reference them in conjunction with logic models in order to provide an
understanding of the context within which the program logic is expected to operate
(Frechtling, 2007). Given their importance in influencing the extent to which a program’s
logic is likely to function as anticipated, it can be extremely helpful to track these kinds
of external nonprogrammatic factors along with the actual performance measures in order
to help interpret and understand the meaning of performance data on a given program
over time or data that are monitored to compare the performance of similar types of
programs or organizations over time. Because these external factors can facilitate or
constrain the performance of public programs and organizations, these kinds of
contextual data can be essential for making sense of performance data, and thus it may be
helpful to build the relevant environmental variables into monitoring systems along with
the actual performance measures.
One special type of external variable that bears mention here is needs indicators,
measures that represent the extent and characteristics of the need for a public program.
While for a public transit agency a principal needs measure might be the percentage of
the population in its service area that is transit dependent, needs indicators for a housing
rehabilitation problem might be the percentage of dwelling units that are not in
compliance with local building and sanitation codes or the percentage of households
living in substandard dwelling units. Similarly, the need for an HIV/AIDS treatment
program might be measured by the percentage of the population living with HIV or the
percentage of new babies born with HIV. The need for a high school student retention
program would likely be measured by the dropout rate before the end of the junior year in
a local school district.
While such needs measures obviously play a crucial role in needs assessments and
program planning efforts, they are germane to program evaluation and performance
monitoring systems as well. Since needs indicators represent the kinds of problematic
conditions that public programs are often designed to remediate, outcome measures often
focus on the extent to which needs have been addressed. For example, one needs
indicator for local police agencies is the number of personal and property crimes reported
per 1,000 population each month. Principal outcome measures for police work include
the percentage of crimes solved and the percentage of reported crimes for which a
conviction is obtained. If police efforts are successful over time in both taking criminals
out of the general population and deterring other potential perpetrators from committing
crimes, we might well expect overall crime rates in the local community to decrease.
Tracking needs indicators, such as baseline crime rates, in ongoing performance
monitoring systems can indeed provide valuable insights into the overall effectiveness of
a program aimed at reducing crime. Baseline crime rates serve as a foundational needs
indicator that establishes the initial level of criminal activity within a community or target
population. This baseline provides a benchmark against which progress and outcomes can
be measured. As the program is implemented, interventions and strategies are designed to
address underlying factors contributing to crime. These may include community policing
initiatives, crime prevention programs, rehabilitation efforts, or social services aimed at
reducing risk factors associated with criminal behavior. The primary outcome of interest
is the reduction in overall crime rates within the program's target area or population. This
outcome reflects the program's effectiveness in achieving its goals of enhancing
community safety, reducing victimization, and improving quality of life.
Measuring the percentage reduction in overall crime rates over time provides a
long-term perspective on the program's impact. It demonstrates the extent to which the
program has successfully mitigated crime and contributed to creating a safer environment
for residents. Needs indicators, such as baseline crime rates, are complementary to
outcome indicators because they contextualize the problem and provide a reference point
for understanding changes in crime patterns. Monitoring both needs indicators and
outcome indicators allows for a comprehensive assessment of the program's overall
effectiveness and its contribution to community well-being. Integrating needs indicators
into ongoing performance monitoring systems ensures continuous tracking of progress
and adjustments as needed. Regular data collection, analysis, and reporting enable
stakeholders to assess trends, identify areas for improvement, and make informed
decisions about resource allocation and program adjustments.
By linking needs indicators with outcome indicators, program managers and
policymakers can engage in evidence-based decision-making processes. Data-driven
insights inform strategic planning, policy development, and resource allocation to
maximize the program's impact on reducing crime and addressing community needs
effectively. Involving community members in the monitoring and evaluation process
enhances transparency, accountability, and community ownership of crime prevention
efforts. Community feedback and perspectives contribute to refining strategies and
fostering sustainable solutions to crime reduction. In conclusion, measuring the
percentage reduction in overall crime rates alongside tracking needs indicators in ongoing
performance monitoring systems provides a comprehensive approach to assessing a
program's effectiveness over time. This integrated approach informs strategic decision-
making, supports continuous improvement, and strengthens efforts to create safer and
more resilient communities. By leveraging data and evidence, programs can demonstrate
measurable impacts and contribute to lasting positive outcomes in crime prevention and
community safety.
D. Integrated Sets of Performance Measures
When a good logic model has been developed for a public program, one can
identify the kinds of measures that would be appropriate to include in a performance
monitoring system directly from the model. It is important to understand that we are not
referring to the actual performance indicators at this point; these will have to be
operationalized in terms of data sources, observations, and definitions regarding what
should count and what should not count, as discussed fully in chapter 4. However, when
we are working with a good logic model, what we should try to measure— the products
and services, knowledge, capabilities, behaviors and actions, conditions and results, and
so forth that constitute performance—should become readily apparent. First, the outputs
and the outcomes shown in the logic model translate directly into output and outcome
measures. Productivity measures will be based on ratios of those outputs to various
categories of resources going to the program, and efficiency measures will be based on
ratios of those same outputs to the dollar value or time invested in producing them.
The outcomes that the classes are intended to produce occur in three strands of
logic. First, teens who complete the program will be knowledgeable about prenatal
nutrition and healthy habits. This will lead to their following proper guidelines regarding
nutrition and health, and this will lead to a higher probability of the desired outcome of
delivering healthy babies. Second, as an initial outcome of the classes, these teens will
also be more knowledgeable about the proper care, feeding, and interaction with their
infants, which will lead to the intermediate outcome of their actually providing the same
to their babies once they are born. Then the delivery of healthy babies and assurance that
they will be provided with proper care should result in these babies achieving appropriate
twelve-month milestones regarding physical, verbal, and social development. The third
strand of logic really leads to a different longerterm outcome: having completed the
classes, these pregnant teenagers are also more knowledgeable about birth control options
and parental responsibilities.
The goal of ensuring responsible sexual behavior among teens, including
abstinence or the use of birth control methods, is crucial for supporting their educational
attainment and readiness for responsible parenting. Programs should prioritize
comprehensive sex education that equips teens with accurate information about
contraception, STI prevention, and the emotional and physical aspects of sexual
relationships. Education empowers teens to make informed decisions about their sexual
health and emphasizes the importance of delaying parenthood until they are emotionally
and financially prepared. Ensuring access to a range of contraceptive methods, including
condoms, oral contraceptives, and long-acting reversible contraceptives (LARCs),
promotes responsible sexual behavior. Accessible contraceptive services reduce barriers
to obtaining birth control and encourage consistent and correct use, thereby reducing the
risk of unintended pregnancies.
Providing counseling and support services that address the emotional, social, and
cultural factors influencing teens' sexual behavior is essential. Counseling sessions can
explore relationship dynamics, communication skills, decision-making processes, and
goal-setting related to education and future family planning. Programs should focus on
building teens' self-efficacy and decision-making skills regarding sexual health.
Empowering teens to assert their boundaries, negotiate safer sex practices, and make
responsible choices contributes to their overall well-being and reduces the likelihood of
unplanned pregnancies. Engaging parents, caregivers, and community members in
supporting responsible sexual behavior among teens is critical. Parental involvement can
enhance communication about sexual health within families, reinforce messages about
delaying parenthood, and provide a supportive environment for teens to navigate
relationships and sexual decisions.
Ongoing monitoring and evaluation of program outcomes, including rates of
contraceptive use, incidence of repeat pregnancies, and educational attainment among
participants, provide feedback for program improvement. Evaluation data guide
adjustments to program strategies and interventions based on evidence of effectiveness.
Encouraging peer support networks and positive role modeling among teens reinforces
responsible sexual behavior norms. Peer-led initiatives, mentoring programs, and youth
advocacy groups promote healthy relationships, respectful communication, and mutual
support in adopting contraceptive practices. By addressing these key components,
programs can effectively promote responsible sexual behavior among teens, reduce the
risk of repeat pregnancies during adolescence, and support educational and life goals.
Emphasizing comprehensive approaches that integrate education, access to contraception,
supportive services, and community engagement creates a supportive environment for
teens to make informed decisions about their sexual health and future aspirations.
Numerous outcome measures are shown because the program logic model shows
three strands of results with multiple outcome stages in each. So the effectiveness
measures range from scores on tests regarding the kind of knowledge the program is
designed to impart, to the percentage of participants delivering healthy babies, the
percentage of these babies achieving appropriate twelve-month developmental
milestones, the percentage of these mothers subsequently reporting abstinence or the use
of recommended birth control techniques, and the percentage of these teenage mothers
who are high school graduates, married, and in no need of public assistance at the time of
their next pregnancy. Cost-effectiveness measures might be defined by relating costs to
any of these outcomes, but the most compelling ones might be the cost per healthy baby
achieving appropriate twelve-month milestones and the cost per repeat premature
pregnancy avoided.
The percentage of teens completing the program who report overall satisfaction
well after their babies have been born can indeed be a meaningful indicator of customer
satisfaction in a teen pregnancy prevention or support program. Satisfaction reported
"well after their babies have been born" provides a long-term perspective on the
program's impact. It reflects whether participants continue to perceive the program
positively over time, indicating sustained satisfaction beyond immediate outcomes. This
indicator goes beyond immediate program completion and captures the broader outcomes
experienced by participants. It assesses whether the program met their expectations and
needs throughout their journey, including during pregnancy and after childbirth.
Satisfaction with the program after childbirth considers participants' overall experience,
encompassing support received, educational content, accessibility of resources, and the
program's relevance to their evolving circumstances as young parents.
High satisfaction levels indicate that the program effectively addresses
participants' needs and provides quality services that are valued by teen parents. It
suggests that the program contributes positively to their lives and contributes to their
overall well-being. Satisfaction serves as a proxy for program effectiveness. High
satisfaction rates often correlate with positive outcomes such as improved knowledge
about parenting, enhanced self-efficacy, better health outcomes for both mothers and
babies, and potentially reduced likelihood of subsequent pregnancies during teenage
years. Conducting satisfaction surveys "well after their babies have been born" ensures
that participants have sufficient time to reflect on their experiences and assess the long-
term impact of the program.
Designing the survey to capture nuanced feedback, including open-ended
questions about specific aspects of the program that were most helpful or areas for
improvement, can provide deeper insights into satisfaction levels. Tracking satisfaction
levels over time allows for monitoring trends and assessing changes in participants'
perceptions as they progress through different stages of parenthood. Incorporating
feedback from program staff, healthcare providers, and other stakeholders who interact
with teen parents can provide a comprehensive understanding of program satisfaction and
areas for enhancement. In summary, the percentage of teens completing the program who
report overall satisfaction well after their babies have been born is a meaningful indicator
that reflects the program's ability to meet the needs of teen parents over the long term. It
underscores the importance of sustained support, program relevance, and positive
outcomes beyond initial participation, highlighting the program's impact on participants'
lives and overall satisfaction with the services provided.
As a second example, consider the logic model proposed for the Canadian
Pension Plan Disability (CPPD) Program shown in figure 3.7. The principal components
of this entitlement program include outreach to potentially eligible recipients of the
program, eligibility determination, client recourse system, case management, and
workforce reintegration. The first four of these components feed into each other in a
pipeline sequence leading to individuals who qualify for the disability benefits applying
to the program, being found eligible for benefits, and receiving monthly benefit checks.
Through periodic assessments of recipients’ capabilities for reentering the workforce and
the provision of vocational rehabilitation services, the fourth program component assists
clients in reintegration into the workforce and leaving the benefit program when feasible.
For each component of the CPPD program, a set of measures has been identified,
including output and outcome measures, along with a mix of efficiency, labor
productivity, service quality, and customer satisfaction measures in order to provide a
balanced portrait of that component’s performance. While the output measures and
various stages of outcome measures are drawn directly from the logic model in figure 3.7,
the efficiency and productivity measures are developed based on ratios of outputs to units
of resources, and the service quality and customer satisfaction measures are intuited from
the model by asking, “What are the most relevant quality dimensions of these outputs and
the service delivery processes associated with them?” and, “What are the most important
aspects of service delivery, outputs, and/or outcomes in terms of customer satisfaction?”
E. Developing Logic Models
Obviously a critical first step in developing performance measures for public and
nonprofit programs is to identify what should be measured. The program logic models
presented in this chapter encourage focusing on end results, the real outcomes that a
program is supposed to generate, and the outputs or immediate products that must be
produced in order to bring about those results. Developing such logic models helps to
identify what is important to measure. But how does one go about developing a logic
model for a particular public or nonprofit program? Looking at formal statements of
mission and, especially, goals and objectives is a good place to begin because they should
articulate the kinds of outcomes that are expected to be produced. Since results-oriented
management systems require performance measures that are directly tied to goals and
objectives, this linkage is discussed in greater detail in chapter 4. Beyond goal statements
regarding outcomes, the logic outlined by a program logic model is based on theory,
experience, and research.
Reviewing relevant academic literature on intervention theories, program plans,
and evaluation studies is crucial for developing a comprehensive logic model. This
process involves gathering and synthesizing existing knowledge and evidence in specific
substantive areas to inform intervention strategies and program design. Engaging
knowledgeable stakeholders in discussions further enriches the development of the logic
model by incorporating diverse perspectives and expertise. Conducting a thorough
literature review helps identify established intervention theories, frameworks, and models
that are relevant to the program's goals and target population. This includes reviewing
scholarly articles, books, research reports, and meta-analyses that provide insights into
effective interventions, causal mechanisms, and factors influencing outcomes.
Examining existing program plans and descriptions offers practical insights into
programmatic activities, intended outcomes, and implementation strategies. These
documents outline the intended scope, resources, timelines, and operational details of the
program, providing foundational information for constructing the logic model. Reviewing
evaluation studies conducted in similar substantive areas provides empirical evidence of
program effectiveness, impact pathways, and lessons learned. Evaluation findings help
validate assumptions, identify best practices, and anticipate potential challenges in
program implementation and outcome attainment. Involving knowledgeable stakeholders
—such as program staff, practitioners, researchers, community representatives, and
funding agencies—in discussions is essential. Stakeholders bring diverse perspectives,
expertise, and contextual knowledge that enrich the development of the logic model.
Their input helps refine program goals, clarify intervention strategies, and ensure
alignment with community needs and priorities.
Based on the literature review and stakeholder input, clearly define the key
components of the logic model: inputs (resources), activities (interventions), outputs
(products or services), outcomes (short-term and long-term changes), and impacts
(broader societal or systemic changes). Map out the causal relationships and logic of how
inputs and activities are expected to lead to desired outcomes and impacts. This causal
pathway should be logical, evidence-based, and supported by empirical findings from the
literature and evaluation studies. Select appropriate indicators and metrics to measure
progress and outcomes at each stage of the logic model. Indicators should be specific,
measurable, achievable, relevant, and time-bound (SMART), facilitating effective
monitoring and evaluation of program performance. The development of a logic model is
an iterative process that may involve revisiting and refining components based on new
evidence, stakeholder feedback, or changes in the external environment. Flexibility and
adaptability are key to ensuring the logic model remains relevant and responsive to
evolving program needs.
Use the logic model as a communication tool to articulate program theory,
demonstrate the rationale behind program design decisions, and foster transparency with
stakeholders, including funders, policymakers, and program beneficiaries. By
systematically integrating academic literature, program plans, evaluation studies, and
stakeholder engagement, organizations can develop a robust logic model that effectively
guides program design, implementation, and evaluation. This structured approach
enhances program effectiveness, promotes evidence-based decision-making, and
facilitates continuous improvement toward achieving desired outcomes and impacts in
specific substantive areas.
There is a wide variety of program logic models in terms of scope and
complexity, level of detail, formatting, and strategies for articulating the flow of logic
leading to outcomes that goes far beyond the scope of this book. The goal should always
be to develop a coherent model that clearly communicates the logic by which
programmatic activities are expected to lead to the intended outcomes. Some books that
provide in-depth discussion of logic models and their development and serve as useful
resources in this regard include McDavid and Hawthorn (2006), Frechtling (2007),
Funnell and Rogers (2011), and Knowlton and Phillips (2012). In addition, the outcome
indicators project has developed prototype logic models and illustrative sets of
performance measures for fourteen specific program areas of interest to nonprofit
organizations and a report published by United Way of America (Hatry, Van Houten,
Plantz, & Greenway, 1996) provides guidelines for developing performance frameworks
and measures for nonprofit organizations.
A backward-mapping approach beginning with identification of longerterm
outcomes is often helpful in delineating the logic underlying program logic models,
especially in program planning but also in performance measurement. With a clear
understanding of what the desired longer-term outcomes are, you can ask, “What are the
impediments to these outcomes occurring on their own, and what has to happen in order
to generate these longer-term outcomes?” Answering these questions will help to identify
the necessary initial and intermediate outcomes that must be generated. They in turn
should tie directly to outputs being directly produced by the program. While beginning at
the outset of the program logic by identifying activities and outputs and then identifying
subsequent outcomes may be effective in developing a logic model, a backward-mapping
approach may at least in some cases lead to a more appropriate logic model because it is
rooted in the desired results from the outset.
Combining the approaches of starting with outcomes and then navigating "back
and forth and up and down" through the various components is fundamental to creating a
robust logic model. A logic model serves as a visual representation that outlines the
logical connections between program inputs, activities, outputs, outcomes, and impacts.
This iterative process ensures clarity in program design, implementation, and evaluation,
aligning activities with intended outcomes and demonstrating the pathways through
which change is expected to occur. The logic model begins by identifying the desired
outcomes or impacts that the program aims to achieve. These outcomes are articulated in
clear, measurable terms that reflect the overarching goals of the program. By defining
outcomes upfront, the logic model establishes a clear vision of what success looks like
and guides subsequent planning and implementation efforts.
Once outcomes are identified, the logic model navigates "back and forth" through
the components to delineate the necessary inputs, activities, and outputs that contribute to
achieving those outcomes. By mapping these components, the logic model establishes the
causal relationships and pathways through which inputs and activities are expected to
lead to desired outputs and outcomes. The iterative process of developing a logic model
fosters clarity and alignment among stakeholders regarding program goals, activities, and
expected outcomes. It facilitates a shared understanding of the program's theory of
change and logic of intervention. By integrating outcomes into program design from the
outset and iterating through program components, organizations can make strategic
decisions about resource allocation, program implementation strategies, and performance
measurement.
Engaging stakeholders throughout the logic model development process ensures
diverse perspectives are considered and promotes buy-in and support for the program.
Logic models should be dynamic documents that can evolve as programs adapt to
changing circumstances, new evidence, or stakeholder feedback. Beyond internal use,
logic models serve as communication tools to articulate program theory, justify resource
needs, and demonstrate accountability to funders, policymakers, and the broader
community. In essence, integrating outcomes-driven planning with a systematic approach
to program design through a logic model enables organizations to articulate their theories
of change, clarify program components, and demonstrate the anticipated pathways to
achieving meaningful outcomes and impacts. This structured approach supports effective
program management, evaluation, and continuous improvement toward achieving desired
results.
Whatever approach is used in a particular performance measurement effort, the
cardinal rule should be never to bend reality to fit a preconceived model. What is
important is to model the program or organization as it is or as it should be; the model
should be thought of as a tool for understanding how the program is intended to operate.
Fortunately, the program logic methodologies presented here are very flexible and should
be adaptable to almost any programmatic or organizational setting. Once the model has
been developed and a strong consensus has been built around it as an appropriate
performance framework, it tends to become the arbiter of issues regarding what aspects
of performance should be included in a monitoring system, and measures of outputs,
quality, efficiency, productivity, effectiveness, cost-effectiveness, and customer
satisfaction can be defined with confidence.
