Logic Models – A Practical Guide

January 2016

Partners in improving local health

1
 Logic Models– Aide Memoire
Introduction
This document has been prepared to assist in the compilation of logic models. It is intended
to provide some background on the theory of logic models and practical help in preparing
logic models.

General comments
A logic model is a roadmap or simplified picture that displays connections between
resources, activities and outcomes within the programme/plan. They are graphic
representations of programmes showing the intended relationships between investments
and results. They are also based around evaluation. It is intended they assist as a model of
change.

"A theory of change is a description of how and why a set of activities-be they part of a
highly focused programme or a comprehensive initiative-are expected to lead to early,
intermediate and longer term outcomes over a specified period." (Anderson, 2000)

A basic logic model in its simplest form:

Inputs Outputs Outcomes

This can be expanded as follows:

Inputs Outputs Outcomes

Programme Activities Short

investments Medium
Long term

1. The summary logic model is described in fig 1.

2. Outcomes answer the “so what?” question. Usually they are short, medium and long
term, relate to learning/action/conditions and are measurable
3. A key question is what is the situation or problem(s) you are trying to solve.
4. Setting of priorities assists in focussing on desired outcomes to ensure activities are
relevant to the problem(s) and impacts.
5. A logic model is just a model. It is a representation of the key issues needing to be
resolved, their potential impacts and intent. Logic models help make our assumptions
explicit.

Logic models do not include all aspects of the programme. They are not intended as
operational plans.

Fig 1

2
Why should you use the logic model? How will it help you?
 Brings detail to broad goals; helps in planning, evaluation, implementation, and
communications
 Helps to identify gaps in our programme logic and clarifies assumptions so success
may be more likely
 Builds understanding and promotes consensus about what the programme is and
how it will work--builds buy-in and teamwork.
 Makes underlying beliefs explicit.
 Helps to clarify what is appropriate to evaluate, and when, so that evaluation
resources are used wisely.
 Summarizes complex programmes to communicate with stakeholders, funders,
audiences.
 Enables effective competition for resources.

The most effective way to develop a logic model is to work backwards and start with
identification of the long-term impacts (fig 2).

Fig 2
Definitions of logic model terminology
Impact - the social, economic, civic and/or environmental consequences of the programme.
Impacts tend to be longer-term and so may be equated with goals. Impacts may be positive,
negative, and/or neutral, intended or unintended.
 Improved health outcomes
 Improved patient experience

Questions about impact once the programme has concluded

 What difference does the programme make?
 Who benefits and how?
 What learning, action, and/or conditions have changed/improved as a result of the
programme?
 At what cost?
 Did we accomplish what we promised? What didn't we accomplish?
 What, if any, are unintended or negative consequences?
 What did we learn?
 What is the net impact?

3
Outcomes - the reason for undertaking the activity and what is achieved when the output is
complete. This includes the results or changes from the programme such as changes in
knowledge, awareness, skills, opinions, aspirations, behaviour, practice, decision-making,
social action or status. Outcomes may be intended and/or unintended: positive and
negative. Outcomes fall along a continuum from immediate (initial; short-term) to
intermediate (medium-term) to final outcomes (long-term). They are measured through
either quantitative or qualitative information.
 Outcomes short term: better connection between services, best practice shared
across health economy, reduction in waste in duplication of effort to resolve x across
health economy
 Outcome medium term: reduction in cases sent outside NCM sites by x%. Increase
in activity undertaken through joint working (measured by cases
 transferred between organisations). Reduction in complaints from patients due to
service delay
 Outcome long term: financial impact of the delivery against the roadmap or similar to
medium term outcomes but with a greater % achievement

Outcomes often fall along a continuum from shorter- to longer-term results. This continuum
is called a "chain of outcomes" (United Way of America, 1996) and the concept - a series of
outcomes that are connected - is fundamental to a logic model.

As you identify your outcomes, think about:

 What might result other than what is intended?
 How else might the programme unfold?
 Who might be affected, unintentionally and/or negatively?
 How might the external environment have unintended influences?

Outputs - what will be achieved by the activities listed the activities, events products, people
reached and participation generated through the investment of resources. Outputs include
such elements as workshops, conferences, counselling, products produced and the
individuals, clients, groups, families, and organizations targeted to be reached by the
activities. Outputs can lead to different outcomes.
 Output: a complete road map for meeting national standards or a complete strategy
for services in the future signed up by all parties

Activities – the tasks or actions undertaken which will produce an output and can produce
more than one.
 Activity: undertake gap analysis or review of current services

Inputs - resources that go into a programme including staff time, materials, money,
equipment, facilities, volunteer time and influencing factors such as policies or other
programmes of activity.
 Input: some national policy or statement about the fit for purpose state of women and
childrens services

Outputs v outcomes
Try not to confuse outcomes with outputs. Outputs are the activities we do or accomplish
that help achieve outcomes. Outcomes are the results of those activities for individuals,
families, groups, or communities. These are described in the following examples:

4
Outputs Outcomes
The programme trains and empowers Community volunteers have knowledge and
community volunteers. skill to work effectively with at-risk youth.
Programme staff teach financial Low-income families are better able to
management skills to low-income families. manage their resources.
The camp experience provides leadership Campers, aged 12-15 years of age, learn
development opportunities for youths. new leadership and communication skills
while at camp.

Logic models can be designed in various formats. An example of a format is shown in fig 3

Logic model
Situation

Impacts

Fig 3
Situation - the originating problem, or issue, set within a complex of socio-political,
environmental and economic circumstances. The situation is the beginning point of logic
model development.

Assumptions – these include the beliefs you have about: the programme, the people
involved, and how you think the programme will work. Assumptions include your ideas
about the problem or situation; the way the programme will operate; what the programme
expects to achieve; how the participants learn and behave, their motivations, etc.; the
resources and staff; the external environment; the knowledge base; and the internal
environment. Faulty assumptions are often the reason for poor results.

External factors - aspects external to the programme that influence the way the programme
operates, and are influenced by the programme. Dynamic systems interactions include the
cultural milieu, biophysical environment, economic structure, housing patterns, demographic
makeup, family circumstances, values, political environment, background and experiences of
participants, media, policies and priorities, etc. Elements that affect the programme over
which there is little control.

When completing your logic model it should be borne in mind that it is not the answer in
itself, as it remains only a model. It is a process that encourages an understanding of the
issues being faced and the various contextual aspects being experienced and a logic model

5
is intended as a road map from problem to solution. It does help you articulate assumed
causal linkages. It does help build consensus in discussion with stakeholders about what
the programme is trying to accomplish. It also helps identify what and when to evaluate.

Secondly, the logic model focuses on expected outcomes. However, there will be
unexpected or unintended outcomes that often occur; either positive, negative or neutral.
Wherever possible you need to think about alternative pathways of change; alternative
outcomes that may occur; and be observant for unintended and unexpected outcomes.

The third limitation that needs to be mentioned is the challenge of causal attribution. A logic
model depicts assumed causal connections and associations; the reasoning behind a
programme; not direct cause and effect relationships. The emphasis is on "reasonable, not
definitive conclusions or absolute proof" (Michael Patton, Utilization-Focused Evaluation,
1997:217). Some people may find this uncomfortable. What actually is attributed to an
effect will vary. There are likely to be many factors that influence observed outcomes.

There are various ways to produce logic models. These are described below:

Approach 1 - work backwards

Start at the end by identifying the long-term outcome(s) of interest. This often results from a
visioning or strategic planning process. Then, work backwards across the model and ask:
1. What preconditions in the medium term must be met for the long-term outcome(s) to
be achieved? You can also phrase the question, "What needs to exist as a precursor
for the long-term outcome(s) to be achieved?"
2. Moving backwards ask what preconditions in the short term must be met in order to
reach the medium-term outcomes? (These are your short-term outcomes.)
3. Who must be involved, reached, targeted, and/or a participant for the short-term
outcomes to be achieved? Be specific about "who" (age, gender, defining
characteristics).
4. What activities, products, events must be undertaken so that those specific
individuals (or groups) will achieve the desired outcomes?
5. Think about: How can these people be reached/engaged? How do they best learn?
Cluster activities into strategies (activities that fit together conceptually) such as
training, media work, coalition development, etc.
6. What resources are needed to conduct these activities, to reach those people, to
effect those outcomes?
7. What assumptions have you made about…? What does research, experience,
wisdom tell us?
8. What external factors outside our control may affect our theory of action?

Approach 2 - focus on activities

1. Programme staff and stakeholders are often most comfortable talking about what
they do in the programme or intend to do…the programme activities. Write down all
activities involved in (or planned for) the programme – workshops, services,
products, etc.
2. For each activity, complete one of the following statements, continue repeating and
completing the statement until your reach a logical end point.
3. “We do __________, SO THAT __________ will occur.”
4. “IF we do__________, THEN__________ will occur.”
5. You can also use the question “But, why?” For example: But, why do I advertise the
workshop? Answer: so that people will attend. But, why? Answer: So that people will
increase their knowledge about…etc.
6. Continue until a chain of connections is created that links programme activities to
desired end results.
7. List the resources needed to ensure the chain of connections is achieved.
8. List all assumptions.
9. List the external factors that may impede your expected theory of action.

6
Approach 3 – start at the end then the front
1. Identify the long-term outcome(s) of interest.
2. Brainstorm all the things that have to happen to reach your long-term outcome(s).
You might have someone record these as the group offers its ideas or each member
can write down his/her own ideas on sticky notes.
3. Using a large workspace, place these items in logical order: what precedes what;
what is connected to what; what is a sequence of what? Check for gaps in the
connections. Keep asking the question "If this, then will this occur?" Use any
metaphor or creative process to capture the connections.
4. You may want to cluster activities into strategies (activities that fit together
conceptually) such as training, media work, coalition development, etc.
5. Gather items that represent assumptions and environmental factors in a special
place and see how they affect your model.
6. Identify resources needed to support the intended theory of action.

Flow of activities
There is merit at the beginning of the logic model process to take a linear approach once you
have confirmed your impacts. By taking an activity and working across the logic model
diagram it is possible then to identify the outputs from that activity and the outcomes (see fig
4). This is a “logical” approach (Flanagan December 2015). Further refinement of logic
model can then proceed to linkages.

Fig 4

Linkages - theory of action

It is the lines and directional arrows in the logic model that provide the depiction of the
connections, or your theory of action. All lines and arrows may be included or abbreviated
and implied. These flows may be vertical and horizontal, one-direction or two-directional,
and show feedback loops (fig 5).

7
It is the linkages, not just what is labelled as input, output, or outcome, that give the model its
validity. Drawing the connections is often messy and time-consuming, but necessary. It is
what helps you make sure you have addressed all the logical connections. Sometimes you
simplify and only include the primary linkages; otherwise, the logic model may become too
difficult to read.

Fig 5

Logic models are not just about inputs, outputs and outcomes that get placed in their
respective columns, rows or boxes. The power of logic models in planning, implementation
and evaluation is how the input, outputs, and outcomes fit together, connect, and relate in
order to achieve desired end results.

Multiple logic models

Multi-level management system
Multiple logic models may be needed to clarify various levels, issues or goals of a single
management system. A national initiative, for example, might include the national (most
macro) level, the state level, and the community level. Each level is depicted with a logic
model in a series of hierarchically linked models. The level of detail may become more
specific as the focus narrows.

These "nested" logic models (Wauchope, 2001; Hernandez, 2000) depict the hierarchy of
various levels and how they connect within a single system. Each logic model is built with
reference to the level above (or below) and in relation to the organization's or programme's
overall mission (fig 6 refers).

Fig 6
Multi-component initiative
In a complex, multifaceted initiative several models might depict the various programmatic
components, goals, sites, or target populations. Each of these "sub logic models" and its
expected outcomes links to the overall logic model to ensure that programmatic outcomes
are achieved. For example, for a community-wide nutrition education programme, there may

8
be one "programme" logic model that provides the "big picture" of the total programme and
then separate, sub logic models for the specific workstreams, projects, components, or
target populations within the community-wide effort. For instance, a community tobacco
control effort might have goals related to youth prevention, clean indoor air, and cessation.
A general logic model depicts the total effort. Separate, more detailed logic models depict
the inputs-outputs-outcomes relative to each component/goal--for example an initiative to
change a restaurant ordinance within the environmental tobacco smoke component.
Expected outcomes for each of the sub-logic models link to the outcomes expressed in the
overall model (fig 7).

Fig 7

Valid logic models and limitations

There are various criteria to be used to assess the validity of a logic model. This includes:
1. Meaningful – it represents action that is valued and worth doing
2. Plausible – it makes sense
3. Doable – it can be carried out
4. Measurable – it can be verified
5. The activities, their outcomes can deliver the impact and address the original
problem(s)

Benefits of logic models

 Helps differentiate between what you do and the results (outcomes)
 Guides and helped focus work
 Increases intent and purpose
 Guides prioritisation and allocation of resources
 Helps to identify important variables to measure

Cautions
 Can become time consuming and/or a paper exercise
 Can become too focussed on outcomes without sufficient spent on the other
elements
 May end up focussing on the wrong thing
 Mixing levels within one logic model
 Not to become fixed and remain flexible

9
How good is your logic model?
Ask yourself:
1. Is each listed outcome truly an 'outcome'? Does the logic model clearly separate
outcomes from outputs, or are the distinctions blurred?
2. Does the highest-level outcome represent a meaningful benefit of value to the
public?
3. Does it have inherent value? Can it be associated with the programme?
4. Is the model truly logical? Do the relationships among the programme elements
make sense and are the casual relationships supported? Three ways to check:
a. Starting at inputs, ask “why?” at each level: why do we need these inputs? Why do
we need to conduct these activities?
b. Starting at the impact level, and working backward, ask “how?” How are we going
to produce these outcomes? The items immediately preceding an outcome should
show “how.”
c. Some elements are necessary but not sufficient. Ask yourself, “What else?”
For example, achieving healthy one-year-olds requires not only achieving a healthy
birth but also achieving proper care during the baby's first year. Asking 'what else?'
helps spot leaps of faith.
5. Are the resources realistic? Is what you intend to do even possible given your
resources?
6. How valid are the assumptions? Are they based on experience and research, or are
they best guesses?
7. Does the logic model reflect the opinions and support of key stakeholders? Were any
stakeholders left out?

Evaluation
The analytical framework which guides the evaluation should be based on a theory of
change with logic models being used to illustrate this. From April 2016, when most
vanguards will start to consider commissioning evaluations, the logic models should be in
place within the vanguard. This should be used as a tool for managing the programme,
engaging local stakeholders in the goals of the programme locally and acting as a skeleton
for the value propositions submitted to the programme team. It should also provide a robust
framework for conducting an evaluation programme. They should:
 identify a set of measures that will be used to establish the success of the
programme;
 identify a source for these data; and
 establish targets which define success.

References
Department of Health
University of Wisconsin, 2008
Enhancing Program Performance with Logic Models, University of Wisconsin-Extension,
Feb. 2003
Adapted from Mike Hendricks, EVALTALK, on-line, 7/9/98; Freddolino, P. 1998, Michigan
Safe and Drug-Free Schools and Communities; Evaluation Training Workshop, Phase 1;
Michigan Department of Community Health.
Strategy Unit, Midlands and Lancashire CSU

NECS Commissioning Support Unit was established in 2012. With staff numbers in
excess of 950 we are based in the north east of England and provide support across the
country to:
- CCGs – contracting, data analysis, CHC, performance management, GPIT, end to
end commissioning
- providers – data management, demand management
- NHS England eg NHS Maternity Services Review
- Vanguards with logic models in urgent care, care homes and acutes as well as a
comprehensive service offerings over the broad range of Vanguard support

10