Personalized System of Instruction (PSI): The Data

PSI may appear to be a new and radical approach to education, but in fact it has been around since the mid-1960s (Keller, 1968). PSI has been used at all levels of education, from elementary schools to university graduate programs, as well as in military and industrial settings (Sherman, 1992). College courses employing PSI have included: introductory psychology, learning, anthropology, sociology, physics, chemistry, economics, business, mathematics, biology, nutrition, psychiatry, library science, home economics, statistics, composition, gerontology, political science, biochemistry, earth science, engineering, and philosophy (Taveggia, 1976; Kulik, Kulik, & Bangert-Drowns, 1990).

Sherman (1992) speculates that PSI has generated a remarkable number of research studies, probably over 2000! What do the data show? Two review articles are especially enlightening. In the first one, Taveggia (1976) concluded "when evaluated by average student performance on course content examinations, the Personalized System of Instruction has proven superior to the conventional teaching methods with which it has been compared" (p. 1032). We might even consider this an understatement, given that PSI was favored in 28 out of 28 comparisons. As Sherman (1992) notes: "This was exciting, particularly because it came from a critic of educational research, who was best known for articles demonstrating that nothing one does in the classroom makes any difference" (p. 59).

A later review by Kulik, Kulik, & Cohen (1979) was even more comprehensive. They reported the following

  • "In 57 of the 61studies comparing PSI and lecture teaching, final examination performance was superior in the PSI section" (p. 310). In 48 of these studies it was possible to express class averages as percentages: "In the typical PSI class...the final examination average was 73.6; in the typical conventional class, the average was 65.9." (p. 311). This is a statistically significant result.
     
  • "A somewhat larger PSI effect---an improvement of about 14 percentage points---is found when achievement examinations are administered several months after the end of the course." (p. 317)
     
  • "Differences between PSI and control classes also tend to be more pronounced on essay than on objective examinations." (p. 317)
     
  • "Differences in student rating of PSI and control classes are also pronounced. Students rate PSI classes as more enjoyable, more demanding, and higher in overall quality and contribution to student learning than conventional classes." (p. 317)
     
  • "The size of the PSI-control differences was also related to the discipline in which a course was offered. But even when PSI-control differences were smallest, a PSI superiority was still quite apparent, and PSI had a statistically demonstrable effect on student achievement. The superiority of personalized instruction was clear under a variety of conditions and with good and poor research designs." (p. 317)

On the downside, a later review by Kulik, Kulik, & Bangert-Drowns (1990) reported:

  • "The percentage of students that complete PSI college classes is smaller than the percentage that complete conventional classes" (p. 286). They were careful to add, however, that better student performance in PSI classes "is not an illusion created by the withdrawal of the weaker students before examination time" (p. 286).

A concern had been raised earlier about critical thinking. Indeed, researchers have also considered this issue. Reboy & Semb (1991) concluded:

  • "First, evidence indicates that PSI has been successfully used to teach complex subject matter at levels of achievement that exceed those attained in a lecture-discussion format. Moreover, studies have demonstrated PSI's effectiveness in teaching higher order cognitive skills. Most of these studies measured transfer effects, but a few specifically demonstrated either gains in or acquisition of higher order cognitive skills." (p. 214)