The Methodology Used To Measure Health Care Consumption During the First Year of the Health Insurance Experiment

Discusses subject matter objectives, measurement theory and pretest results used to formulate a design for measuring the use of health services. An overview of the questionnaires, incentives, and procedures used by the Health Insurance Survey in its first year in the field is given, and a preliminary evaluation of response is presented. Attention is focused on the possible existence of a reporting bias correlated with experimental treatment that, if undetected, could bias estimates of demand. Pretest data suggested this kind of bias could be present so steps were taken to minimize it in the year one measurement design. A preliminary evaluation suggests the magnitude of the correlated bias was reduced but possibly not eliminated entirely. This contributed to the decisions to drop the control group from the study, to introduce additional reporting incentives for some groups, and to expand the study design to provide estimates of any remaining reporting biases.