TIPS FOR LOGIC MODELS AND RESULTS FRAMEWORKS Logic models1 and results frameworks (such as the logical framework) contain a series of “results statements” - short descriptions of outcomes etc. - that projects, programmes or organisations expect to achieve or contribute to. These models and frameworks Always include one or more levels of outcome Usually include an impact, goal or ultimate outcome Will include activities2 and outputs if they represent projects; but often not if they represent organisations or long-term programmes. The results statements are arranged in logical, sequential, relationships with one another. Results at one level (e.g. short-term outcomes) are expected to happen before those at the next level (e.g. intermediate outcomes) - and to be necessary conditions for them to happen. In logic models, the results are presented two-dimensionally - horizontally and vertically (see Figure 1). Figure 1 Logic model In a results framework, they are presented vertically (see Figure 2). This makes room for further columns of information. The results column is often called the narrative summary. Figure 2 Results framework Results 1 Indicators MoVs Assumptions Logic models are sometimes described as Theories of Change (ToCs). ToCs however – expressed either as narrative or diagrammatically - often have wider scope than logic models and include for example the problems that need to be addressed and a wider range of assumptions and other external factors. They are also not always presented in a linear fashion, and may include feedback loops or be presented as systems. 2 Activities, and in some jurisdictions outputs, are not classified as “results”. Logic models are more visual and generally easier to comprehend than results frameworks, but they contain less information. Results frameworks contain indicators and means of verification (data sources and data instruments), and often also baseline and target values for the indicators (in which case they are usually called performance frameworks because they define particular performance standards). Both logic models and results frameworks should contain information on assumptions or risks relating to the achievement of results at the different levels. Tips for results statements 1. The results should be important – they should represent resolutions of significant problems or take-up of significant opportunities. A sequence of situation, problem and objectives analyses is often used to identify important results. 2. The results should be legitimate and appropriate for your organisation to pursue, bearing in mind its mandate, experience and expertise. Strategy analysis often follows situation, problem and objectives analyses to identify the particular results that your organisation should focus on. 3. Results should be expressed as results not activities, describing the expected change or completed product or service. “To promote more accountability in government” → “Government more accountable”. ”Training for accountants” → “Accountants trained” 4. They should clearly identify the people or organisation(s) that will be involved in the change or will receive/experience the product or service. Avoid imprecise terms like “stakeholders” and “partners”. 5. The results should be evaluable – capable of being verified. This is usually assured by the combination of points 3 and 4 above. Results statements do not however have to be directly measurable, unlike indicators. 6. Don’t blur the line between outputs and outcomes. Outputs are what the project itself can deliver. Outcomes are what happens as a result of a combination of what the project delivers and external factors and conditions. 7. Results statements should be as simple as possible. Ideally there should be only one dimension. The following multi-dimensional statement should probably be split into two or even three statements “Increased accountability, transparency, and civil society participation in decision making in local government”. 8. If there are two or more elements, they should be achievable at a similar stage of the project or its aftermath and should not be sequential. Avoid statements that contain phrases like “X leading to Y”; or “Y as a result of X”. This creates ambiguity about what the evaluable result is. Tips for logic models or narrative summaries 1. The different levels of results should be logical and represent a clear progression. Results at the lower level should logically happen before the higher level and contribute to it. (See Figure 3.) A higher level result should be more than the sum of the results at the lower level – it should also represent a change which builds on, and takes forward the lower level results. 2. Progression between the levels should be plausible. Results at one level – together with the stated assumptions – should make a convincing case for the related result at the higher level to happen. If the progression from the results at one level to the next “stretches the imagination” it may mean that you have too few levels in your model. You may need to insert another level of outcomes. Alternatively, it may mean that your project has unrealistic higher outcomes and you should be more modest in you ambition, choosing outcomes at every level that you can plausibly make some contribution to. 3. The necessary outputs should be achievable by the project, programme, etc. in the timeframe, given the inputs (resources, expertise, partnerships, etc.) and what is known about the operating environment. 4. Logic models, as well as performance frameworks, should ideally include key assumptions. Tips for assumptions 1. Assumptions can be expressed as risks if this seems more logical to you. 2. Assumptions should be beyond the direct control of your intervention. If they are within your field of control – such as having an adequate budget or set of competencies in the project team - they are likely to be pre-conditions that you need to fix or have strategies for fixing before the intervention starts. 3. Only include significant assumptions or risks – those you need to monitor and have contingency plans for. On the other hand, if you identify “killer” risks - those that are likely to happen and which will seriously reduce your chance of success if they do you need to take your project back to the drawing board. Tips for indicators and means of verification 1. Indicators should have a clear and precise “unit of measurement”. For example: % of smallholder clients of project-supported financial service providers who perceive that new products or practices effectively meet their financial needs 2. Indicators should represent the core characteristic(s) of the result in question and be at the same level in the results chain. For example, if the result is about changed behaviour, the indicator should not be about attitudes or intentions. 3. An indicator needs to be linked to a pre-identified data source, and a defined instrument for collecting the data. The data need to be accessible and reliable, and data collection needs to be timely and affordable. 4. Gender and other relevant categories of disaggregated data should be available where needed. 5. Indicators usually need targets. But try to avoid defining targets until you have enough of an insight to define them realistically. Targets often depend on the prior identification of baselines. Patrick Spaven October 2014