Set the Stage for your Evaluation


Preparing for an evaluation is a lot like preparing for other important activities. You have to determine what information you need, lay out a plan, and make some decisions about how you will proceed.

Gather Background Information

Doing interviews, reviewing reports, searching in dusty archives—whatever it takes—read everything you can to learn about all the aspects and the various nuances of your project. (Escape route: This step is only necessary if you’re starting from the beginning as an evaluator. If you already have a program, initiative, or program in place, then you’ve got the information you need.)

The background information you need is often in the hands of other people—especially if you are an external evaluator— someone who is conducting an evaluation of a project that he or she did not design.

If you have not been integrally involved with the design of the project or program, you should read everything, from reports to memoranda and meeting notes. In fact, review any documentation that helps you better understand the project. And don’t forget to talk to the program designers and other key stakeholders. They’re an invaluable source of information that can help you better understand the project. They will also help shape questions, identify credible sources, and provide encouragement and critical feedback. Finally, they will render the incalculable service of helping you review and interpret your findings.

Together, the information you pull together from these sources will help you develop a conceptual framework or logic model. Many program designers find it very useful to do this, because then you can view at a glance—and discuss with ease—the project’s motivation and intentions, components, strategies, and desired outcomes.

Develop a Logic Model

A Logic Model is used to conceptualize a single intervention, project, initiative, or program. While it is helpful to build one to understand the relationships among the implementation steps of a project, program, or initiative and the intended outcomes and impact of those activities, it works best if all the activities fall within one initiative. For example, if the goal is to implement a new literacy program in a school, then a logic model can depict the activities that will be undertaken to get that literacy program in place.

For a more detailed explanation of logic models, click here. For more information about developing a logic model, go to Lab 1: Logic Model.

Design Plan

In preparing to collect data, it’s a good idea to consider the use of tools that will help you organize evaluation questions, constructs, and methods. Using organizing tools helps ensure that you’ve touched on as many key aspects of the project as is necessary to develop a complete and reliable picture. And it forces you to consider the many different methods for capturing these aspects.

Constructs and Indicators

One way to make sure your evaluation plan is comprehensive and touches on all the related components and activities is to develop a set of constructs (i.e., major topics or areas, concepts or big ideas) and indicators (i.e., measures of success that will help to answer your questions).

To learn more about constructs and indicators, click here.

Question/Method Matrix

A useful approach is to develop a matrix with the questions along one side and the methods along the top. Then you can check boxes on the grid to determine which methods will provide answers to which questions. It’s good to have more than one method for each question so that you can “triangulate” your data—or use multiple sources of evidence to support your findings.

To see a matrix that shows how to outline key constructs, list questions, and plot them across a wide range of methods, click here.

Data Needs

In many cases, you will want to collect data from multiple respondent groups using multiple strategies. This will allow you to “triangulate” the findings. That is, they allow you to use multiple sources of information to support a particular finding. For instance, in evaluating an instructional program, student test scores, performance on classroom work, and teacher reports all have the potential of supporting a finding of improved student achievement. So when you organize the targets and methods of the evaluation, you help ensure that multiple sources will lead to triangulation of the data.

Consider each of your evaluation questions to determine what data you will need to answer them:

  • Student achievement data?
  • Teacher perspectives on instructional programs?
  • Parent reactions to family events?
  • Longitudinal data on achievement by grade level?
  • Observations of classroom practices?
Once you know your needs, it’s easier to identify the methods you will use to collect the data.
Select Methods

The next step in designing the evaluation is determining how you will go about measuring your constructs and answering your research questions. You need to select the right set of methods to answer the questions you have specified.Tape Measure

After you’ve determined the questions you hope to answer (and perhaps the constructs you will need to measure), you can begin to consider how best to find the data you need to answer those questions. You can collect data in a variety of ways. You can design surveys or questionnaires, conduct interviews, hold focus groups, or combine them together to capture more information. These types of data collection methods fall into two general categories: quantitative and qualitative methods. Which one is right to monitor your implementation, or should you use a combination of methods? These days it is very common to use the latter—referred to as a “mixed methods” approach because it provides the richest information about how things are going.

Analyze Data
Analysis of
Qualitative Data

Analyze DataYou’ve developed an evaluation design. You’ve determined the best means to collect your data, and then you’ve done it—collected all of it. Now you’re at the point of making sense of all the data available to you. But how do you do it?

Analyses of both qualitative and quantitative data can yield a rich pool of information, but pulling it out of the raw data requires that you follow a few basic steps—carefully. Presenting your findings in a clear and convincing way is the final step in this phase of your evaluation.

What will you do with the information?

This is where the whole evaluation process was leading from the beginning. What decisions can you make based on the data collected? What actions can you—should you—take? This is where you decide what changes you and others want to make to improve the program or what steps to take to initiate a new one. Recall the questions you identified in the beginning.
How did you answer them?
What did you learn?
How will you use this information?
What did you find?

In some ways, this is the most satisfying stage of your evaluation. After preparing for and designing your evaluation, after collecting and analyzing your data, you can now tell the world (or at least the people important to your project) what you found.

ReportA good evaluation report is clear, concise, and provides adequate evidence for claims and enough explanation to make sure the reader understands your interpretation of the data. It is sometimes tempting to include too much information. When you have collected stacks and stack of surveys and reams of field notes, it is difficult to know “when to say when.” You’ve become invested in each of your data tables, but if the data don’t show anything, leave them out. Be clear about how the evaluation was conducted as well. Everyone involved in developing the school improvement plan or who is affected by the school improvement plan will be interested both in the methods and the outcomes of the evaluation.

Taking action means implementing specific strategies to accomplish your goals—to “put the rubber on the road,” as they say. Develop an action plan based on the data collected. Changes or improvements may focus on the content of the program, format, delivery, staffing, follow-up strategies, activities, setting, resources, and on and on. It all depends on what your data tell you. And all the decisions do not need to be made at one point in time. Collecting data to make course corrections should occur in an iterative way—one change leading to another after the results of the first change are assessed.

Tips on Making Decisions and Taking Actions
  • Consider whether you think the findings are valid for your program. Validate the data by looking for support for one set of data in another set. Do the findings clearly apply to your situation?
  • Determine what actions/decisions are suggested by the findings. Focus on areas to address, but you don’t try to address everything at once.
  • Determine whether possible actions are feasible. The data may suggest changes that are not really possible for you to make—given resources, time, or other constraints.
  • You may need to do additional research or information-gathering on particular strategies or program adjustments. Don’t jump into something without knowing enough about it to know whether it is likely to work for your set of circumstances.
  • Determine how you will know whether the changes/improvements are working. Put a monitoring plan in place that will allow you to watch implementation carefully. Don’t forge ahead without examining how well things are going as you proceed.