Skip navigation

In the context of grants, evaluation is a systematic inquiry into project performance. In its formative mode, it looks at what is working and what is not; in its summative mode, it looks at what did work and what did not. In both modes, it identifies obstacles to things working well and suggests ways to overcome them. For evaluation to proceed, the events or conditions that it looks at must exist, must be describable and measurable, and must be taking place or have taken place. Its focus is actualities, not possibilities.

 

Data Collection:

Effective evaluation requires considerable planning. Its feasibility depends on access to data. Among the more important questions to consider in collecting data for evaluation are:

  • What kinds of data need to be acquired?
  • What will be the sources of data?
  • How will sources of data be sampled?
  • How will data be collected?
  • When and how often will the data be collected?
  • How will outcomes with and without a project be compared?
  • How will the data be analyzed?

 

Problem Definition:

In developing an evaluation plan, it is wise to start from the problem definition and the assessment of needs and work forward through the objectives to the evaluation methods. After all, how a problem is defined has inevitable implications for what kinds of data one must collect, the sources of data, the analyses one must do to try to answer an evaluation question, and the conclusions one can draw from the evidence.

 

Evaluations pose three kinds of questions: descriptive, normative, and impact (or cause and effect). Descriptive evaluation states what is or what has been. Normative evaluation states what is to what should be or what was to what should have been. Impact evaluation states the extent to which observed outcomes are attributable to what is being done or has been done. The options available for developing an evaluation plan vary with each kind of question.

 

Power:

An evaluation plan does not need to be complex in order to provide useful answers to the questions it poses. The power of an evaluation should be equated neither with its complexity nor with the extent it manipulates data statistically. A powerful evaluation uses analytical methods that fit the question posed; offer evidence to support the answer reached; rule out competing evidence; and identify modes of analysis, methods, and assumptions. Its utility is a function of the context of each question, its cost and time constraints, its design, the technical merits of its data collection and analysis, and the quality of its reporting of findings.

 

Constraints:

Among the most common constraints on conducting evaluations are: time, costs, expertise, location, and facilities. Of these constraints, time, costs, and expertise in particular serve to delimit the scope and feasibility of various possible evaluation design options.

 

Design Options:

Most evaluation plans adopt one of three design options: experimental, quasi-experimental (or non-equivalent comparison group), or pre/post In the context of observed outcomes, the experimental option, under random assignment of participants, is most able to attribute causes to outcomes; the pre/post option – even one featuring interrupted time series analyses – is least able to make such attributions.

 

The experimental option tends to be the most complex and costliest to implement; the pre/post option tends to be the simplest and least costly. Increasingly, Federal grant programs favor the experimental evaluation design, even in areas of inquiry where it is costly and difficult to implement at the necessary scale, such as education and social services.

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: