Kurt and colleagues (2004) compared the computer and pencil-and-paper versions of self-report instruments for depression and ability to function. Many elderly people (over 65) experience depression, but it is not often addressed in primary care settings. A computerized measure which could be completed by a patient while waiting would be of great benefit to both patients and clinicians.
The instrument being studied was the revised Center for Epidemiologic Studies of Depression Scale, or CESD-R20. A total of 34 patients completed all baseline data and follow-up retest a week later. Correlations were computed for computer and pencil versions and for tests over time. Most of the correlations were 0.70 or above, with the exception of the correlations between the CESDR20 computer and paper at follow-up (0.605), and between the baseline / follow-up for the paper CESDR20 (0.615).
The authors concluded that this preliminary study was successful, and that patients expressed favorable attitudes toward the computerized measure, even though 72% of the patients had no experience with computers. They advised larger studies in future research, as well as asking nonparticipants their reasons for refusal, since this could be an unknown factor that might change the results.
This was an essential study for the field of psychological measurements because it addressed the effects of a changeover from paper to electronic forms. The transition from paper to computer is seen in every part of health care, and it is essential to know whether computer versions of tests are sufficiently reliable for previous data on paper tests to be useful. If not, the psychometric properties of the electronic form must be determined.
This study advances the material in the textbook in two main ways. The authors measured test-retest reliability as well as parallel forms reliability. According to Kaplan & Saccuzzo (2013), test-retest reliability may not be valid in measuring factors that change over time, such as level of depression. This is probably why the authors waited only a week before the retest. However, testing again so soon brings up the potential for carryover effects (Kaplan & Saccuzzo, 2013, p.109). It is uncertain why the authors used test-retest reliability at all, considering these potential issues.
The study also used the parallel forms method. The computer-assisted and pencil-and-paper versions were identical except for the difference in form, so the expectation was that each form would yield the same results. In fact, the results were very similar but not exactly the same. Psychologists may want to explore further the relationship between computer and paper versions to ensure continuity of records.
The study also recognized that computer-assisted testing can allow doctors and nurses to carry out other tasks during the assessment, the test can be given without bias, and it can be self-paced so that there is no time limit that might skew a person’s responses (Kaplan & Saccuzzo, p. 195). With regard to the standards for educational and psychological testing, this research added to the body of knowledge concerning reliability of depression screeners for the elderly. Reliability and validity are extremely important in psychological testing, and if an instrument does not have both to a reasonable degree, it should not be used.