Thirty-six non-hospitalized subjects with chronic pain from OA of the knee participated in an evaluation of transcutaneous electrical nerve stimulation (TENS) and naproxen, an NSAID. All pre-experiment treatment was withdrawn. Each subject experienced in some order three 3-week treatment phases: NSAID plus placebo TENS; TENS plus placebo drug; and double placebo. A broad range of pain measures was used, including daily diary ratings, and four-times-per-day ratings entered into a small electronic data logger (the PIPER) worn by the subject. A substantial placebo response occurred across all conditions, which may have masked treatment differences. Broad comparisons across subjects, combining the four main measures of pain, found no significant differences among the three experimental treatments. Analysis of diary and PIPER data for individuals suggested that, in a small minority of subjects, the NSAID plus placebo TENS combination may be more effective than double placebo. The PIPER ratings seemed to tap aspects of the pain experience different from those captured by conventional measures, suggesting the value of very frequent pain assessments, such as those entered by a subject into the PIPER, in the study of chronic pain.