Biostatistical computing

Abstract

The purposes of data analysis are the purposes of research itself—to make new discoveries and to establish their certainty to the extent that they may be entered into the body of firm scientific knowledge that is the foundation for future research and for the application of what we have learned. From this, two conclusions follow: First, data analysis should involve the researcher himself, albeit with the help of biostatisticians and other methodological experts. Second, data analysis for the sake of discovery should be valued as highly as data analysis for the validation of hypotheses. Already computers have brought to biomedical research unprecedented support for full utilization of hard-won experimental data, and thev doubtless have played an important role in the increasing standards of quality for scientific reporting. But it is possible that the most exciting developments lie ahead, where new interactive techniques will increasingly involve the biomedical researcher himself and will contribute to evolution of the scientific method by providing innovative supports for the process of hypothesis discovery.