SCALPEL: Prototype for a computer assisted finite language for data screening

Abstract

No one has yet been able to devise a computerized method for using a natural language to analyze large data files. As a result scientists and physicians have been limited to using record forms and questionnaires for their information collection. On large studies where computer assistance is required the researcher may lose the opportunity to communicate directly with his own data. Many times he is forced to communicate with a computer programmer/analyst, who then operates on his data. We feel that a finite language based on a questionnaire format is an excellent means of allowing the researcher to interrogate his own computerized data. We have created SCALPEL as a prototype language system for multiple statistical inquiries for the non-computer oriented medical scientist; particularily the epidemiologist. The data used to demonstrate this finite language is from a national study of obese persons.