Because the science of nursing is still at an early stage of development, nurses involved in clinical research often find that no appropriate measures are available to sample the theoretical construct of interest. Researchers often must design and test new instruments before proceeding to test theory. A study of the effect of nurse-client contracting on family planning knowledge and behaviors in a university student population required the development of a new instrument. For nursing research, it was necessary to obtain pre- and post-test data about the extent of client knowledge about family planning. For clinical nursing assessment, it was essential to make a complete and accurate appraisal of the client's knowledge about family planning. Test items were written and rated by three expert judges for face validity and for content validity. Item revision resulted in an 86-item instrument on human fertility, reproduction, and contraception. The instrument was field-tested with 38 female university students. While the resulting test-retest reliabilty was adequate, problems arose in examining construct validity. By reducing the instrument to 46 items to meet the requirements for scientific rigor, important client data would be missing from the assessment process. A 76-item compromise instrument was developed which contained the 46 items in a construct on knowledge about family planning. The final instrument was complete, accurate, and appropriate for research. This instrument development represents one situation in which conflicting values emerge as part of the clinical nursing research process. (NB)