Skip to Main Content
The promise of breakthroughs in computer security - experimental test beds, insider-detection advancements, biometrics, and user interfaces that are robust to human error-will remain empty as long as methodological details trail the hype. A few selected issues - operational definitions, reliability, internal validity, and external validity - serve to illustrate common sources of experimental error that may be brought to one's attention by a written methodology. The first two issues regard valid measurement, and the second two regard valid experimentation. Together, they exemplify the importance of strong methodology in pushing the envelope of security.
Date of Publication: March-April 2005