Office of Justice Programs Office of Justice Programs
BJA Center for Research Partnerships and Program Evaluation
BJA Center for Research Partnerships and Program Evaluation

Characteristics of Good Performance Measures

Performance measures permit the systematic assessment of the degree of success a program has had in achieving its stated objectives and activities. Good performance measures have the following characteristics:

  1. Objective - free from bias or without distortion by personal feelings, prejudices, or interpretations

    For example: One method of assessing learning may be to survey the attendees of training to ask them how much they learned at the training. A more objective approach would be to administer a pre- and post-test asking a series of questions on the topics addressed in the training (e.g., what is the correct definition of an 'objective'? What are the two different types of performance measures?) and to compare the scores. This change in scores provides objective information on how much was learned.

  2. Relevant - applicable to the objectives and activities

    For example: If the program's activities are intended to reduce recidivism by improving the skills of program participants, there should be a measure that captures changes in these skills.

  3. Clear - unambiguous and easy to understand

    For example: If the measure requires calculations or interpretation, it should include instructions on how these calculations are to be performed or how something should be interpreted or defined.

  4. Reliable - yields consistent, stable, and uniform results over repeated measurements

    For example: If a scale is used to measure attitudes, it should yield consistent results over time (assuming that what is being measured has not changed).

  5. Valid - accurately measures what it is supposed to measure

    For example: Research has shown that people have difficulty accurately recalling details of arrest history if events occurred several months ago or if the history is extensive. For a measure that includes the date of arrest, an official arrest report is likely a more valid data source than self-reported criminal histories.

  6. Practical - the knowledge, time, personnel, and ability to implement are available

    For example: If the measure requires Department of Corrections data, those responsible for reporting should have access to these data and the ability to gather and report these data.

  7. Allows for gradation of success

    For example: It is common for participants of substance abuse treatment to relapse. Rather than simply counting those who successfully complete treatment or those who relapse, it may be more useful to measure something that allows for gradations of success, such as number of relapses or number of dirty urines.

  8. Appropriate for the clients of the program

    For example: If a program serves children, the measure should use language children will understand.

  9. Verifiable - the reported information can be confirmed

    For example: The number of reported convictions is verifiable because court records can be used to ensure that the reported data are correct.

  10. Creating a performance measure that meets every one of these criteria perfectly may be difficult; however, these criteria should at least be considered when selecting performance measures.

    NEXT >