Assessing recalibration as a response-shift phenomenon.

Background: Traditionally, the difference between pretest and posttest scores is used as an estimate of change. This can be problematic when repeated self-report measures are used to assess change resulting from interventions intended to change beliefs, behaviors, attitudes, or values about health or safety. If the intervention is effective, participants may apply more stringent criteria in response to a posttest questionnaire than they did at the pretest. This kind of change has been termed a recalibration response shift.

Objectives: To present scale recalibration as a measurable response shift, and to illustrate a method that can be used to estimate its magnitude and direction: the retrospective pretest.

Methods: In a quasi-experimental study investigating the effectiveness in small construction companies of narrative simulation exercises targeting back and fall injuries, a retrospective pretest was administered concurrently with a delayed posttest 4 months after the simulation exercises.

Results: : In the first intervention year, the results from a brief (two-item) retrospective pretest pertaining to safety climate were consistent with a recalibration response shift in the intervention group, but not in the control group. In the second intervention year, when all 10 items of the safety climate questionnaire were used for the retrospective pretest, no evidence of recalibration was found.

Conclusions: Although the evidence of recalibration was equivocal, the findings illustrate circumstances in which recalibration response shifts may occur and characteristic patterns of findings may suggest that recalibration has or has not occurred.