In March 1998, OMB granted approval to test the wording and format of materials designed to collect aggregate `race' and `Hispanic origin' data according to new standards. The first step in the testing process was to develop three different forms, each approaching the problem of aggregate reporting in a slightly different way. The second step was to test the three forms to determine their strengths and weaknesses and, ultimately, to select a single form to present to OMB. The first wave of testing was conducted in cognitive labs at the Bureau of Labor Statistics (BLS) and the National Center for Health Statistics (NCHS) by professional staff from 3 different agencies: BLS, NCHS, and the U.S. General Accounting Office (GAO). After assessing the information collected in the lab about the three forms, the next step was to take the form into the field for cognitive testing in the workplace. This paper presents the results of those two waves of testing. It also considers the differences and similarities in the results obtained from cognitive testing in formally different settings and reviews the possible sources of those differences and similarities in findings.