Abstract:

The swift adaptation of digital computer programs and languages to public opinion poll data processing has revolutionized the working methods of market research firms. However, while the data processing methods have undergone radical changes, the fact remains, nonetheless, that "field" data collection and their coding into a language processable by electromechanical or electronic data processing systems have not followed along and are presently still relatively "archaic". Taking the example of data transcription into a computer exploitable language, various methods have already been employed: "the transfer sheet" , the "Mark Sensing and magnetic cards, etc...", "the precoded questionnaire", etc... None of these methods yields satisfactory results either in terms of transcription accuracy, time limits or cost requirements. We have in the course of the past three years, together with MARPLAN France and IBM France, been the first in the world to develop an infrastructure for the collection and checking of rough field data directly by electronic data processing systems "without human intervention", at a heretofore unequalled accuracy, rapidity and cost.