Carroll et al.: “The Minimal Manual” Precise

About the Study

Carroll et al. developed a Minimal Manual for a word processing program in order to address problems in the training process that occur with standard self-instruction manual. They applied Minimalist training principles in the design of the manual and tested the result in two experiments. Based on these experiments, the researchers concluded that the Minimal Manual helped users perform better and more efficiently than the standard self-instruction manual.

Designing the Minimal Manual

Principles of the Minimal training model: Carroll et al. explain that their “strategy in training design was to accommodate, indeed to try to capitalize on, manifest learning styles and strategies” (74). Based on prior research on the subject, they outline the principles of the Minimal training model. These principles include focusing on real world activities and tasks, cutting out information that users often overlook, supporting error recognition and recovery, and providing “guided exploration” (75-77).

Design Process: Carroll et al. argue that Minimalist design is similar to other aspects of user interface design in that it is “developed iteratively: designed, empirically evaluated, and then redesigned” (77). They explain how the principles of the Minimal training model influenced their design (77-81), as well as the procedures they used to test and revise their document prior to the two experiments (81-84).

Experiment 1

Purpose: The purpose of this experiment was to “contrast a commercially developed standard self-instruction manual (SS) with the experimentally developed Minimal Manual (MM) in an office-like environment” (84).

Participants: Nineteen subjects were chosen by an outside agency. Ten of them participated in the MM condition while the other nine participated in the SS condition (85). These subjects were screened to have experience with typical office work but little experience with word processing software. None of them had prior experience with the software used in the study (86).

Procedure: This experiment was a “between-subjects contrast of the independent variable of manual (MM or SS)” (84). Within each condition, subjects were placed into groups of two or three in a simulated office environment. They were given the training manual for their condition and asked to finish prerequisite training and perform a related task. This process was repeated eight times and covered different material. As the experiment was supposed to mimic a real environment, subjects were allowed to talk to each other, utilize the system library, and call a support hotline in addition to referring to their manuals. The researchers collected two dependent measures: the time to complete the training and performance tasks and the performance on the eight performance tasks (86-88).

Results and Discussion: The researchers discuss a number of statistically significant results regarding time and task performance. They found that users of the Minimal Manual accomplished more tasks in less time than their SS counterparts. The researchers conclude that these results are strong indicators that the Minimal Manual has a better design than the standard self-instruction manual, but they could not determine why it was better from this experiment.

Experiment 2

Purpose: The purpose of this experiment was to “the contrast between a commercially developed, standard self- instruction manual (SS) and the experimentally developed Minimal Manual (MM).” (89).

Participants: Thirty-two subjects were chosen by an outside agency. Eight of them participated in each condition (MM, SS, “learn while doing” (LWD), and “learn by the book” (LBB)) (89). These subjects were screened for the same qualities as the first experiment. Three participants were replaced due to frustration (91).

Procedure: The experiment had a “2x2 between-subjects design” (90). The LWD participants received five hours to perform tasks, while the LBB participants received three hours to “use the manual in order to learn about the system” (89-90). They separately received two more hours to perform other tasks. They were encouraged to use the manual to perform tasks, but the entire library was available. The researcher sat with each participant, and the participants were encouraged to think aloud as they completed six tasks. Researchers measured time and performance as in Experiment 1, but also measured attention and effort involved (92). They achieved this by coding participants’ actions and tabulating errors.

Results and Discussion: This experiment also yielded statistically significant results. Measures that were used to analyze learning and targeted errors and skills indicated that MM subjects “performed better and more efficiently” (97).

General Discussion

When comparing the two experiments, the researchers conclude that the two “converge on the conclusion that the Minimal Manual is substantially and reliably superior to the commercial self-instruction manual” (99). They point out that this study can not be used to generalize to “other areas of educational technology” (99).