The critical-incident technique (CIT) is one of many types of job-analysis procedures. The CIT is often used because it produces behavioral statements that make explicit to an employee what is required and to a rater what the basis for an evaluation should be.

The following steps are involved in a complete CIT procedure:

1. Identify the major dimensions of job performance This can be done by asking a group of raters and ratees to brainstorm and generate dimensions relevant to job performance. Each person lists, say, three dimensions. The group members then combine their lists and eliminate redundancies.

2. Generate “critical incidents” of performance For each dimension, the group members should list as many incidents as they can think of that represent effective, average, and ineffective performance levels. Each person should think back over the past 6 to 12 months for examples of performance-related behaviors that they have witnessed. Each incident should include the surrounding circumstances or situation.

   If you are having trouble generating incidents, you might want to think of the following situation: Suppose someone said that person A, who you feel is the most effective person in the job, is a poor performer. What incidents of person A’s behavior would you cite to change the critic’s opinion?

   Try to make sure that the incidents you list are observable behaviors and not personality characteristics (traits).

3. Double-check that the incidents represent one dimension This step is called retranslation. Here you are trying to make sure there is clear agreement on which incidents represent which performance dimension. If there is substantial disagreement among group members, this incident may need to be clarified. Alternatively, another dimension may need to be added or some dimensions may need to be merged.

   In the retranslation process, each person in the group is asked to indicate what dimension each incident represents. If everyone agrees, the group moves on to the next incident. Any incidents on which there is disagreement are put to the side for further examination at the end of the process. At that time they may be discarded or rewritten.

4. Assign effectiveness to each incident Effectiveness values are assigned to all the incidents that survived retranslation. How much is incident “A” worth in our organization, on, say, an effectiveness scale of 1 (unacceptable) to 7 (excellent)? All group members should rate each incident. If there is substantial disagreement regarding the value of a certain behavior, that behavior should be discarded.

The chart on the following page shows some CIT worksheets for you to try your hand at. The dimensions included are a subset of those generated in a research project conducted for a hospital that wanted a common evaluation tool for all non-nursing employees.*

The jobs covered ranged from floor sweeper and clerical worker to laboratory technician and social worker. Of course, the behavioral standards for each dimension differed across jobs—an excellent floor sweeper behavior would not be the same as an excellent lab technician behavior. The dimensions included in the worksheets appear fairly generic, though, and are probably applicable to jobs in most organizations. You may want to develop more specific dimensions or other dimensions altogether.

Remember, after generating incidents, your group should determine agreement levels for the dimension and value for each incident. An easy way to do this is for one person to recite an incident and have everyone respond with dimension and value. This process could be informal and verbal or formal and written.