s08_synthesis_logic_models.qmd

---
title: Logic Models and Synthesis Development
---

:::{.callout-learning}

By the end of this session, you will be able to:

- Describe the structure of Logic Models and how to interpret them
- Explain how logic model components (inputs, outputs, outcomes) can be reframed to be more relevant for a synthesis research collaboration
- Apply the principles of Logic Models to the development and refinement of synthesis research questions
:::

## Logic Models

Logic models are a planning tool that are designed to support program development by depicting the flow of resources and processes leading to a desired result. They are also used for outcomes-based evaluation of a program and are often requested as part of an evaluation planning process by funders or stakeholders.

### Inputs, Outputs, Outcomes

A simplified logic models comprise three main parts: Inputs, Outputs and Outcomes.

![](images/synthesis_logic_models/LM1.png){fig-alt="A diagram of three boxes lined horizontally, labeled 'inputs,' 'outputs,' and 'outcomes' from left to right" width="70%" fig-align="center" .lightbox}

Inputs reflect ***what is invested***, outputs are ***what is done*** and outcomes are the ***results of the program***.

In a more detailed logic model, outputs and outcomes are further broken down:

* Outputs are often represented as 'Activities' and 'Participants'. By including participation (or participants), the logic model is explicitly considering the intended audience, or stakeholders, impacted by the program. Engagement of this audience is an output. 
* In the case of outcomes, these can be split into short, medium and long-term outcomes. Some people may label this last category 'Impact.'

![](images/synthesis_logic_models/LM2.png){fig-alt="The three-box model above is expanded to include resources under inputs, activities and participants under outputs, and short, medium, long term under outcomes." width="70%" fig-align="center" .lightbox}

Defining the inputs, outputs and outcomes early in a planning process enables teams to visualize the workflow from activity to results and can help mitigate potential challenges. Logic models can be thought of as having an "**IF** this **THEN** that" structure where inputs {{< fa arrow-right >}} outputs {{< fa arrow-right >}} outcomes.

![](images/synthesis_logic_models/LM3.png){fig-alt="The figure above is modified to include gray arrows pointing right to left between each box, with each box labeled 'if' and each arrow labeled 'then', for example IF resource condition is met THEN activities can result, IF activities THEN participants, etc" width="70%" fig-align="center" .lightbox}

In the example below we have constructed a simple logic model for a hypothetical project where training materials are being developed for a group of educators to implement at their respective institutions.

![](images/synthesis_logic_models/LM4.png){fig-alt="The figure above is modified to include specifics for each box: resources: 'We have sufficient money, cycles, expertise', outputs: 'We can develop online data science curricula for educators', outcomes: short term, 'Educators can access material and learn about data science,' medium term, 'They can implement instruction in data science fundamentals', and long term 'Widespread data literacy across broad populations of students'" width="70%" fig-align="center" .lightbox}

### Linkages

Linkages are not always sequential and can be within categories, bi-directional and/or include feedback loops. Detailing this complexity of relationships, or theory of action, can be time consuming but is a valuable part of the thought process for project planning. In exploring all relationships, logic modeling also allows for assessing program feasibility.

![](images/synthesis_logic_models/LM5.png){fig-alt="Multiple boxes, each representing an input, an output, or an outcome, are connected by gray arrows flowing mostly left-to-right but some are bi-directional or indicate loops back to previous steps" width="70%" fig-align="center" .lightbox}

:::{.callout-tip}
## Other variations

The above graphics include two sections within Outputs - Activities and Participants - and this is quite common. There is variation in logic model templates, including versions with a third type of output - "Products'. Sometimes description of these products is contained within the Activities section - for example, 'develop curricula', 'produce a report' - however calling these out explicitly is beneficial for teams focused on product development.
:::

### Situations, Assumptions, External Factors

Program development (and logic modeling) occurs in response to a given 'Situation' or need, and exploring this is the first step in modeling. The situation defines the objective, or problem, that the program is designed to solve hence some logic models may omit the left-hand situation column but be framed with Problem and Solution statements. Finally, comprehensive logic modeling takes into consideration assumptions that are made with respect to the resources available, the people involved, or the way the program will work and also recognizes that there are external factors that can impact the program's success.

![](images/synthesis_logic_models/LM6.png){fig-alt="A blank template for a logic model, with boxes for inputs (resources), outputs (activities, participants, products), and outcomes (short, medium, long term); on the left side of the entire diagram is a box labeled 'situation'; below the diagram are two boxes labeled 'assumptions' and 'external factors'" width="70%" fig-align="center" .lightbox}

:::{.callout-note}
## In Summary

Logic models support program development and evaluation and comprise three primary steps in the workflow:

-   **Inputs:** Resources, contributions, and investments required for a program;
-   **Outputs:** Activities conducted, participants reached, and products produced; and
-   **Outcomes:** Results or expected changes arising from the program structured as short-, medium- and long-term.

:::

## Logic Models for Synthesis Development

Logic models are one tool for program development and have sufficient flexibility for a variety of situations, including planning for a research collaboration. While some logic model categories may feel less relevant (can we scale up to a long-term outcome from a published synthesis?), the process of articulating the research objective, proposed outcome, associated resources and activities has value. Below are examples of questions that a typical logic model (**LM**) will ask, and how these might be reframed for a research collaboration (**RC**).

:::{.callout-note title="Objective/Problem Statement" icon=false}
<!--blue border to match graphic-->

**LM:** What is the problem? Why is this a problem? Who does this impact?

**RC:** What is the current state of knowledge? What gaps exists in understanding? Why is more information or synthesis important?
:::

:::{.callout-caution title="Inputs" icon=false}
<!--yellow/orange border to sorta match graphic-->


**LM:** What resources are needed for the program? Personnel, money, time, equipment, partnerships...

**RC:** What is needed to undertake the synthesis research? For personnel, think in terms of the roles that are needed - data manager, statistician, writer, editor etc. Consider the time frame. DATA - what data are needed and what already exists?
:::

:::{.callout-tip title="Outputs" icon=false}
<!--green border to sorta match graphic-->

::::{.panel-tabset}
## Activities

**LM:** What will be done? Development, design, workshops, conferences, counseling, outreach...

**RC:** What activities are needed to conduct the research? This could be high level or it could be broken down into details such as the types of statistical approaches.

## Participants

**LM:** Who will we reach? Clients, Participants, Customers...

**RC:** Who is the target audience? Who will be impacted by this work? Who is positioned to leverage this work?

## Products

**LM:** What will you create? Publications, websites, media communications...

**RC:** What research products are planned / expected? Consider this in relation to the intended audience. Is a peer-reviewed publication, report or white paper most appropriate? How will derived data be handled? Will documentation, workflows, or code be published?
::::
<!--end tabset-->
:::

:::{.callout-important title="Outcomes" icon=false}
<!--red border to sorta match graphic-->

:::{.panel-tabset}
## Short-term

**LM:** What short-term outcomes are anticipated among participants. These can include changes in awareness, knowledge, skills, attitudes, opinions and intent.

**RC:** Will this work represent a significant contribution to current understanding?

## Medium-term

**LM:** What medium-term outcomes are predicted among participants? These might include changes in behaviors, decision-making and actions.

**RC:** Will this work promote increased research activity or open new avenues of inquiry?

## Long-term Outcomes

**LM:** What long-term benefits, or impacts, are expected? Changes in social, economic, civic, and environmental conditions?

**RC:** Will this work result in local, regional or national policy change? What will be the long-term impact of increased investment in the ecosystem?
::::
<!--end tabset-->
:::


## Tools

### Slide Template

A good starting point for drafting a logic model is a [logic model template](resources/synthesis_logic_models/logic_model_template.pptx)

* And here is the same [logic model template on Google Drive](https://docs.google.com/presentation/d/1MBUy35QIAkA5FNCdRzWJM2aPaFXfUmGl/edit?usp=sharing&ouid=113328660159760492310&rtpof=true&sd=true){.external target="_blank"}
    
<!--image is screenshot but links to the Google Doc-->
[![](images/synthesis_logic_models/ppt_template.png){fig-alt="A blank logic model template in PowerPoint, with boxes for inputs (resources), outputs (activities, participants, products), and outcomes (short, medium, long term)" width="50%" fig-align="center"}](https://docs.google.com/presentation/d/1MBUy35QIAkA5FNCdRzWJM2aPaFXfUmGl/edit?usp=sharing&ouid=113328660159760492310&rtpof=true&sd=true){.external target="_blank"}

### Mermaid Flowchart

[Mermaid flowcharts](https://mermaid.js.org/syntax/flowchart.html){.external target="_blank"} are supported in [Quarto documents](https://quarto.org/docs/authoring/diagrams.html#mermaid){.external target="_blank"}.  Mermaid is a JavaScript-based tool for creating diagrams and flowcharts using text-based syntax.

Here is a simple example of a logic model Mermaid flow chart for a hypothetical synthesis project.

:::{.panel-tabset}
## Flow chart diagram

```{mermaid}
flowchart LR
    INPUTS --> ACTIVITIES --> OUTPUTS --> OUTCOMES/IMPACTS

    Scenario{{Accelerate synthesis via data science training}}

    R1[Instructor] & R2[Classroom space] & R3[Projector] --> B{Data Science Workshop}
    B --> C(Workshop Curriculum)
    B --> D(Presentations and Practice)
    
    C & D --> E[/Improved Delta management/] & F[/Increased analytic efficiency/]
```

## Source code

To create the diagram, create a Mermaid code block in a Quarto document (note the `{mermaid}` in the header) and use [Mermaid syntax](https://mermaid.js.org/syntax/flowchart.html){.external target="_blank"} to define the flowchart.

```{{mermaid}}
flowchart LR
    INPUTS --> ACTIVITIES --> OUTPUTS --> OUTCOMES/IMPACTS

    Scenario{{Accelerate synthesis via data science training}}

    R1[Instructor] & R2[Classroom space] & R3[Projector] --> B{Data Science Workshop}
    B --> C(Workshop Curriculum)
    B --> D(Presentations and Practice)
    
    C & D --> E[/Improved Delta management/] & F[/Increased analytic efficiency/]
```
:::

## Breakout Exercise

:::{.callout-exercise}
## Synthesis Planning

In breakout groups, focus on refining ideas for synthesis topics using the logic modeling tools described in this section. 

**Step 1: Logic Model Template**

As a breakout group, download (pptx) or make a copy (Google slides) of the logic model template and begin to flesh out a high-level logic model:

* Summarize the synthesis challenge
* Define the inputs needed to approach the synthesis
* Define the outputs, including activities and products that would would be needed to address the issue
* Define the short term outcomes and longer-term impacts of the work

Often it is helpful to start with a brainstorming activity to list activities and products that might be used to address the synthesis challenge, then connect those in terms of outcomes and impacts, and then circle back to the resource and data inputs needed to feed the logic model. Thinking of the whole model as a workflow can help conceptualize the dependencies among steps.

**Step 2: Logic Model Workflow**

Once you have outlined a logic model, consider the workflow from inputs to outcomes. Are there any gaps in the flow? Are there any assumptions that need to be made explicit? Are there any external factors that could impact the success of the project?  

Map out on paper, or using the Mermaid flowchart syntax, a workflow from inputs to outcomes. This can help identify potential challenges and areas where additional resources or planning may be needed.
:::