Logic Modeling
Logic Modeling
:
:
A Tool to Guide Program Design
A Tool to Guide Program Design
and Performance Measurement and
and Performance Measurement and
Evaluation: Telling your
Evaluation: Telling your
Performance Story
Performance Story
John A. McLaughlin
John A. McLaughlin
[email protected]
My Aim Today
My Aim Today
Change the way you think about
Change the way you think about
conceptualizing and telling your
conceptualizing and telling your
performance story.
performance story.
Change the way you think about
Change the way you think about
Performance Measurement and Evaluation
Performance Measurement and Evaluation
related to your program.
Two Questions
Two Questions
What is the right
What is the right
Outcome
Outcome
?
?
–
– ShortShort--term term –– produce researchproduce research –
– Intermediate Intermediate –– develop prototype; market develop prototype; market penetration
penetration
–
– Strategic Strategic –– impact on clientsimpact on clients
Am I getting at the right Outcome, the
Am I getting at the right Outcome, the
right
right
way
way
? Working better! Management
? Working better! Management
Control
Control
– – EfficiencyEfficiency – – EffectivenessEffectivenessDrivers for Program Planning, Performance
Drivers for Program Planning, Performance
Measurement, and Evaluation
Measurement, and Evaluation
Government Performance Results Act (GPRA)Government Performance Results Act (GPRA)
–
– Requires Strategic Plan, Annual Plan, and Annual Requires Strategic Plan, Annual Plan, and Annual Performance Report
Performance Report
OMBOMB’’s Program Assessment Rating Tool (PART)s Program Assessment Rating Tool (PART)
–
– Tool designed to assess and evaluate programs across Tool designed to assess and evaluate programs across the government
the government
DD ActDD Act
–
PART Purposes
PART Purposes
"The program assessment effort presents an opportunity
"The program assessment effort presents an opportunity
to inform and improve agency GPRA plans and reports,
to inform and improve agency GPRA plans and reports,
and establish a meaningful systematic link between
and establish a meaningful systematic link between
GPRA and the budget process.
GPRA and the budget process.
The PART is a diagnostic tool that relies on objective
The PART is a diagnostic tool that relies on objective
data to inform evidence
data to inform evidence--based judgments to assess and based judgments to assess and evaluate programs across the wide range of issues
evaluate programs across the wide range of issues
related to performance. . . . The formalization of
related to performance. . . . The formalization of
performance evaluation through this process is intended
performance evaluation through this process is intended
to develop defensible and consistent ratings of programs
to develop defensible and consistent ratings of programs
for the FY 2004 Budget and beyond."
PART Sections
PART Sections
Purpose/Relevance/ Purpose/Relevance/ Federal Role Federal RoleTo assess whether the program design
To assess whether the program design
and purpose are clear and defensible
and purpose are clear and defensible
Strategic Planning
Strategic Planning To assess whether the agency sets To assess whether the agency sets valid annual and long
valid annual and long--term goals for term goals for the program the program Program Program Management Management
To rate agency management of the
To rate agency management of the
program, including financial oversight
program, including financial oversight
and program improvement efforts
and program improvement efforts
Program Results
Program Results To rate program performance on goals To rate program performance on goals reviewed in the strategic planning section
reviewed in the strategic planning section
and through other evaluations
Orientation
Orientation
Words we
Words we
’
’
ll use a lot!
ll use a lot!
–
–
A
A
program
program
is
is
…
…
–
–
Performance Measurement
Performance Measurement
is
is
…
…
–
–
Performance Measurement leads to
Performance Measurement leads to
…
…
–
–
Program Evaluation
Program Evaluation
is
is
…
…
–
–
Program Evaluation leads to
Program Evaluation leads to
…
…
What
What
’
’
s common among all?
s common among all?
–
More Words
More Words
GoalsGoals ObjectivesObjectives – – ProgramProgram ShortShort--termterm
IntermediateIntermediate – – SupportingSupporting ResourcesResources ActivitiesActivities
Orientations for Performance Measurement
Orientations for Performance Measurement
& Evaluation
& Evaluation
PERFORMANCE MEASUREMENTPERFORMANCE MEASUREMENT
–
– AccountabilityAccountability
What objectives/outcomes have been accomplished at what What objectives/outcomes have been accomplished at what levels?
levels?
PROGRAM EVALUATIONPROGRAM EVALUATION
–
– Learning, Program ImprovementLearning, Program Improvement, , DefenseDefense
What factors, internally and/or externally influenced my What factors, internally and/or externally influenced my performance? (
performance? (RetrospectiveRetrospective))
What effect will this level of performance have on future What effect will this level of performance have on future performance if I don
performance if I don’’t do something? (t do something? (ProspectiveProspective))
Orientation of Data Needs
Orientation of Data Needs
Horizontal
Horizontal
–– Information used within Information used within ‘‘levellevel’’ to manage for to manage for results. All this information should be useful to
results. All this information should be useful to
managers within level; some will ALSO be
managers within level; some will ALSO be
useful to managers at other
useful to managers at other ‘‘levelslevels’’..
Vertical
Vertical
–– Information collected at various Information collected at various ‘‘levelslevels’’ that is that is shared across levels to enable organizational
shared across levels to enable organizational
managing for results and accountability.
Evaluation Axioms
Evaluation Axioms
Evaluation follows the Evaluation follows the plan, do, monitor, and refine plan, do, monitor, and refine
process.
process.
(Even evaluators don
(Even evaluators don’’t get things right the first t get things right the first time out!)
time out!)
You should never evaluate a program that you do You should never evaluate a program that you do not know.
not know.
–
Logic Models as Recipes
Logic Models as Recipes
Recipes have three essential components:
Recipes have three essential components:
–
– A description of the entree to be produced; A description of the entree to be produced; –
– A list of specific ingredients according to A list of specific ingredients according to specific measures;
specific measures;
–
– and specific steps to put the ingredients and specific steps to put the ingredients together.
together.
A good cook follows the recipe
A good cook follows the recipe
–
–
managers
managers
would do well to create and follow their
would do well to create and follow their
recipe for success!
Recipes and Logic Models
LOGIC MODEL
LOGIC MODEL
Strategic Aim Intermediate outcome Short term outcome Customers Outputs Activities Resources WHY HOWPROGRAM STRUCTURE OUTCOME STRUCTURE
EXTERNAL CONDITIONS INFLUENCING PERFORMANCE (+/-)
A logic model is a diagram and text that illustrates/describes the
logical (causal) relationships among program elements and the
results to be achieved
Sphere of Influence
Elements of the Logic Model
Elements of the Logic Model
Resources/Inputs:Resources/Inputs: Programmatic investments available Programmatic investments available to support the program.
to support the program.
Activities:Activities: Things you doThings you do–– activities you plan to conduct activities you plan to conduct in your program.
in your program.
Outputs:Outputs: Product or service delivery/implementation Product or service delivery/implementation targets you aim to produce.
targets you aim to produce.
Customer: Customer: User of the products/services. Target User of the products/services. Target audience the program is designed to reach.
audience the program is designed to reach.
Outcomes:Outcomes: Changes or benefits resulting from activities Changes or benefits resulting from activities and outputs.
and outputs.
Outcome StructureOutcome Structure
–
– ShortShort--term (term (K, S, AK, S, A) ) –– Changes in learning, Changes in learning, knowledge, attitude, skills, understanding
knowledge, attitude, skills, understanding ((ObjObj.).) –
– Intermediate (Intermediate (BehaviorBehavior) ) –– Changes in behavior, Changes in behavior, practice or decisions
practice or decisions ((ObjObj.).) –
– LongLong--term (term (ConditionCondition) ) –– Changes in condition Changes in condition (Goal)(Goal)
External Influences:External Influences: Factors that will influence change Factors that will influence change in the affected community.
Benefits of Logic Modeling
Benefits of Logic Modeling
Communicates the performance story of the
Communicates the performance story of the
program or project.
program or project.
Focuses attention on the most important
Focuses attention on the most important
connections between actions and results.
connections between actions and results.
Builds a common understanding among
Builds a common understanding among
staff and with stakeholders.
staff and with stakeholders.
Helps staff
Helps staff
“
“
manage for results
manage for results
”
”
and informs
and informs
program design.
program design.
Finds
Finds
“
“
gaps
gaps
”
”
in the logic of a program and
in the logic of a program and
works to resolve them.
Steps in the Logic Model Process
Steps in the Logic Model Process
1
1 Establish a stakeholder work group and collect Establish a stakeholder work group and collect documents.
documents.
2
2 Define the problem and context for the program or Define the problem and context for the program or project.
project.
3
3 Define the elements of the program in a table.Define the elements of the program in a table. 4
4 Verify the logic table with stakeholders.Verify the logic table with stakeholders. 5
5 Develop a diagram and text describing logical Develop a diagram and text describing logical relationships.
relationships.
6
6 Verify the Logic Model with stakeholders.Verify the Logic Model with stakeholders. 7
7 Then use the Logic Model to identify and confirm Then use the Logic Model to identify and confirm
performance measures, and in planning, conducting
performance measures, and in planning, conducting
and reporting performance measurement and
and reporting performance measurement and
evaluation.
Step 1. Establish a stakeholder work group
Step 1. Establish a stakeholder work group
and collect documents and information.
and collect documents and information.
Convene/consult a stakeholder work group
Convene/consult a stakeholder work group
–
– provides different perspectives and knowledge provides different perspectives and knowledge –
– attempts agreement on program performance attempts agreement on program performance expectations
expectations
Review sources of program or project Review sources of program or project documentation
documentation
–
– Strategic and operational plansStrategic and operational plans –
– Budget requestsBudget requests –
– Current metricsCurrent metrics –
– Past evaluationsPast evaluations
Step 2. Define the problem the
Step 2. Define the problem the
program addresses and the
program addresses and the
context.
context.
The Context Drivers of Success Constraints on Success The Program Factors leading to the Problem 1 2 3* your program niche The Problem the Program AddressesStep 4. Verify the logic table with
Step 4. Verify the logic table with
Stakeholders.
Stakeholders.
Outcomes
External Influences: Budget is reduced for next FY.
Long-term (Changes in Condition) Evaluation culture established. Quality of evaluations managed and conducted is improved. Improvements in effectiveness of programs leading to better environmental results Intermediate (Changes in Behavior) Number of evaluations conducted and managed increased. Program evaluation skills are used by customers in the work environment. Short-terms (Changes in K,S,A) Knowledge of program PM/PM increased. Customers equipped with skills to manage and conduct evaluations. Customers Reached NCEI Staff IAC Staff PEN PEC 04’ OSWER OW States HQ/Regional managers & staff Outputs PE /PM Training Materials. EPA Managers and staff complete training. Innovation Training Materials. Activities Develop training materials. Deliver Training. Resources/ Inputs 5 ESD Staff $65K Extramural funds Outside Consultant
Step 5. Develop a diagram and text
describing logical relationships.
Draw arrows to indicate/link the causal
relationships between the logic
model elements.
– Limit the number of arrows. Show only the most critical feedback loops.
– Work from both directions (right-to-left and left-to-right)
– There are many different forms of logic model diagrams. We use these resources … We use these resources … For these activities … For these activities … To produce these outputs. To produce these outputs.
So that the customers can change these ways.
So that the customers can change these ways. Which leads to these outcomes. Which leads to these outcomes. Leading to these results! Leading to these results! Outcomes Indicators Impacts
Step 6. Verify logic with
Step 6. Verify logic with
stakeholders.
stakeholders.
Seek review from the same, or an even
Seek review from the same, or an even
broader, group of stakeholders.
broader, group of stakeholders.
Check the logic
Check the logic
-
-
again
again
–
– HowHow--Why Questions. Start with Outcomes and Why Questions. Start with Outcomes and ask
ask ““How?How?”” Start at Activities, ask Start at Activities, ask ““Why?Why?”” –
– IfIf--Then Questions. Start at Activities and move Then Questions. Start at Activities and move along to Outcomes asking
along to Outcomes asking ““If this, then that?If this, then that?””
Compare to what units in the organization
Compare to what units in the organization
do and define their contributions to the
do and define their contributions to the
outcomes.
outcomes.
Check the logic by checking it against
Check the logic by checking it against
“
“
Z
Z
”
”
Logic
Logic
Unpacking the Program
Unpacking the Program
’
’
s Logic
s Logic
In real life programIn real life program’’s achieve their strategic results s achieve their strategic results through a series of actions similar to a relay race.
through a series of actions similar to a relay race.
–
– Action A produces a set of outcomes that become Action A produces a set of outcomes that become inputs to Action B.
inputs to Action B.
–
– Action B produces a set of outcomes that become Action B produces a set of outcomes that become inputs to Action C.
inputs to Action C.
–
– Action C produces a set of outcomes that lead to the Action C produces a set of outcomes that lead to the final strategic goal of the program.
final strategic goal of the program.
These actions could be thought of as nested These actions could be thought of as nested programs within the larger program.
Action A Resources A Outcomes Action B Resources B Outcomes Action C Resources C Outcomes Strategic Program Results!
“
“
Z
Z
”
”
Logic
Logic
Supplier
Supplier
-
-
Customer Relationship
Customer Relationship
Unpacking supports more focused PM and thus more useful evaluation, as well as better understanding and
communication about how the “Program” is supposed to work!
Program $,Staff Perform Research Ideas for technology change for Industry researchers Leads to applications in energy technologies Potential for technology change documented Develop Technology Lab Prototype Report for Users and Manufacturers leads to Commercial Prototype Technology Available for Commercialization Commercial $, Staff Produce technology & Educate Market Manufacture the Technology in market for buyers of that technology leads to Technology accepted, purchased Consequences of use - lower energy costs and emissions Competitive economy, cleaner environment
External Influences: Price of oil and electricity, economic growth in industry and in general, perception of risk of global climate change and need for national energy security, market assumptions, technology assumptions
Deploy Technology Policies, Incentives, Information for Users and Manufacturers leads to knowledge, less risk perceived Early Adopters express desire to buy Customers Outcomes-Resources
and Partners Activities Outputs
for Customers Reached Short-term Outcomes Intermediate Outcomes from Long-Term -Problem Solution (shared responsibility)
‘
‘
Z
Z
’
’
Logic Model for an Energy R,D,&D
Logic Model for an Energy R,D,&D
Program
Program
Source: McLaughlin and Jordan, 1999
Added Resources
Added Resources
Improving Water Quality Training Program Resources * EPA * State * Local * Private Clean/ Safe/ Swimable /Fishable Water
Level I Logic Model
Resources * EPA * State * Local * Private Materials Development Recruitment Training Technical Assistance Website Developers/ Builders * Increased awareness of harmful effects * increased awareness of new technologies and incentives Developers and builders acquire new technologies and change practice Reduction in NPS pollutants in waterways Healthier Wetlands More Fish Clean Beaches Clean/ Safe/ Swimable /Fishable Water
Level II Logic Model
Level III Logic Model
Resources * EPA * State * Local * Private Materials Development Recruitment Training Technical assistance Website Materials Ready Trainees Ready Developers/ Builders Trained Trainees Receive TA Trainees/ Others aware of/ using new information Developers and builders acquire new technologies and change practice Reduction in NPS pollutants in waterways Healthier Wetlands More Fish Clean Beaches Clean/ Safe/ Swimable /Fishable Water * Increased awareness of harmful effects * increased awareness of new technologies and incentives
Complex Effects Chain
Complex Effects Chain
PartnersShared Common Outcomes Transparency
Step 3. Define the elements of the
Step 3. Define the elements of the
program or project in a table.
program or project in a table.
- HOW
-External Influences:
Resources/
Inputs Activities Outputs Customers Short-term Intermediate Long-term
Reached (Changes in K, S, A) (Changes in Behavior) (Changes in Condition)
Key Questions Managers Need to
Key Questions Managers Need to
Answer about their programs
Answer about their programs
What am I doing, with whom, to whom/what? (What am I doing, with whom, to whom/what? (efforteffort))
How well am I doing it? (How well am I doing it? (qualityquality))
–
– Customer FeedbackCustomer Feedback –
– Peer Review for Technical QualityPeer Review for Technical Quality –
– User Review for Social ValidityUser Review for Social Validity
Is anybody (anything) better off? (Is anybody (anything) better off? (effecteffect))
–
– ShortShort--termterm –
– LongLong--termterm
What role, if any, did my program play in the results?What role, if any, did my program play in the results?
What role, if any, did the context play?What role, if any, did the context play?
Were there any unintended outcomes?Were there any unintended outcomes?
What will happen if I donWhat will happen if I don’’t do something?t do something?
PM
Evidence?
Evidence?
Evaluation aims at Evaluation aims at testing hypothesestesting hypotheses::
–
– If I expose these people to this intervention, within this If I expose these people to this intervention, within this context, I
context, I’’ll get these results.ll get these results.
Three buckets of information are required:Three buckets of information are required:
–
– Intervention; Results (shortIntervention; Results (short--term, Intermediate, and term, Intermediate, and long
long--term term –– IMPACT); and ContextIMPACT); and Context..
Logic ModelsLogic Models enable the selection of appropriate enable the selection of appropriate measures to acquire the information required.
Pro gra m Lo gic Hi era rch y Hie ra rch y of P er for m an ce M ea su rem en t D at a
Program Logic Elements Matching Levels of Performance Information
7. End results
6. Practice and behavior change
5. Knowledge, attitude, and skill changes
4. Reactions
3. Participation
2. Activities
1. Resources
7. Measures of impact on overall problem, ultimate goals, side effects, social and economic consequences
1. Resources expanded; number and types of staff involved; time extended
2. Implementation data on what the program actually offers or does
3. The characteristics of program
participants and clients; numbers, nature of involvement; background
4. What participants and clients say about the program; satisfaction; interest; strengths; weaknesses
5. Measures of individual and group changes in knowledge, attitude, and skills
6. Measures of adoption of new practices and behavior over time
Types of Performance Measures
Types of Performance Measures
Category
Category DefinitionDefinition ExamplesExamples
Resources/ Resources/ Inputs
Inputs
Resources consumed by the Resources consumed by the organization.
organization.
Amount of funds, # of FTE, materials, Amount of funds, # of FTE, materials, equipment, supplies (etc.).
equipment, supplies (etc.). Activities
Activities The work performed that directly The work performed that directly produces the core products and produces the core products and services.
services.
# of training classes offered as # of training classes offered as
designed, according to standards of designed, according to standards of excellence.
excellence. Outputs
Outputs Products and services Products and services
delivered/produced as a direct result of delivered/produced as a direct result of program activities.
program activities.
Research and evaluation reports Research and evaluation reports produced. produced. Customer Customer Reached Reached
Measure of target population receiving Measure of target population receiving outputs.
outputs.
% of target population trained; # of % of target population trained; # of target population receiving technical target population receiving technical assistance. assistance. Customer Customer Satisfaction Satisfaction
Measure of satisfaction with outputs.
Measure of satisfaction with outputs. % of customers dissatisfied with % of customers dissatisfied with training; % of customers
training; % of customers ““very very satisfied
satisfied”” with assistance received.with assistance received. Outcomes
Outcomes Accomplishment of program objectives; Accomplishment of program objectives; attributable
attributable to program outputs.to program outputs.
Pounds of pollutants reduced; Miles of Pounds of pollutants reduced; Miles of beaches cleaned; % increase in
beaches cleaned; % increase in industry
industry’’s understanding of regulatory s understanding of regulatory recycling exclusion; # of sectors that recycling exclusion; # of sectors that adopt regulatory recycling exclusion. adopt regulatory recycling exclusion.
Work Quality Measures
Work Quality Measures
Efficiency
Efficiency The ratio of the amount of input to the The ratio of the amount of input to the amount of output. Focus is on
amount of output. Focus is on operating efficiency. Relating operating efficiency. Relating
output to some specific resource in output to some specific resource in terms of cost or time.
terms of cost or time.
Cost per workbook Cost per workbook
produced; Cost per produced; Cost per inspection conducted. inspection conducted.
Productivity
Productivity Measure of the rate of production per Measure of the rate of production per some specific unit of resource (e.g. some specific unit of resource (e.g. staff or employee). The focus is on staff or employee). The focus is on labor productivity.
labor productivity.
Number of enforcement Number of enforcement
cases investigated per cases investigated per inspector. inspector. Cost Cost Effectiveness Effectiveness
Measure that relates outcomes to costs.
Measure that relates outcomes to costs. Cost per pounds of Cost per pounds of
pollutants produced; pollutants produced; Cost per mile of beach Cost per mile of beach cleaned. cleaned. Service Service Quality Quality
Measure of the quality of products and Measure of the quality of products and
services produced. services produced. Percent of technical Percent of technical assistance requests assistance requests
responded to within one responded to within one week.
Types of Data
Types of Data
Qualitative
Qualitative
–– Observations, interviews, document reviews, Observations, interviews, document reviews, “
“descriptions of incidents, actions, processesdescriptions of incidents, actions, processes””
Quantitative
Quantitative
–– NumericalNumerical---- data collection through reports, data collection through reports, tests, surveys, existing data bases
tests, surveys, existing data bases
Not a question of either/or, but when to use
Not a question of either/or, but when to use
a method given performance question and
a method given performance question and
context. Using
context. Using
mixed
mixed
methods yields the
methods yields the
strongest conclusions.
In the end Logic Models:
In the end Logic Models:
Enable the identification of outcomes for which Enable the identification of outcomes for which you are
you are responsibleresponsible and those for which your are and those for which your are
accountable accountable..
Focus your Focus your PM/PEPM/PE on the right elements of on the right elements of performance to enable
performance to enable program improvementprogram improvement and and the estimation of
the estimation of causal relationshipscausal relationships between and between and among elements.
among elements.
Better position you to Better position you to present and defend your present and defend your claims.