Assembly Training Experiment #2

Assembly Training Experiment #2

Experiment I showed that retention of an assembly skill could be more effectively trained with VE’s than with other media. We hypothesized that this was primarily because of the interactivity afforded by VE’s. Experiment II was designed to clarify the effect of interactivity and to replicate the previous findings using a more powerful within-subjects design.

Methods


Participants. 24 people (12 women) were recruited from the University of Washington Psychology Department’s Human subject pool. Participants were given extra credit for their introductory Psychology class in return for their participation.
Materials. The real world tasks on which all participants were tested involved correctly placing eight blocks into a correct configuration in each of three ‘circuit board’ puzzles. Each participant learned how to solve three different boards – one in each training medium. Pilot testing showed that the three boards’ solutions were equally difficult to learn, however, the boards differed in block color and overall shape. (Color and shape differences were used to enhance the memorability of each board solution.)

Three environments were used to train people how to perform this task: a video tutorial, an interactive VE trainer, and a non-interactive VE trainer. A great deal of effort was devoted to making the video tutorial effective and complete. The interactive VE was identical to that used in experiment one. The non-interactive VE consisted of a video tape of a previous participant’s interactive VE session.

Prior to their training, participants were given a block rotation test (similar to Vandenberg’s) developed by Marcy Lansman. The test requires users to identify whether two perspective drawings represent the same three dimensional block. Participants complete as many problems as they can in four minutes. The percent correct on this test was used as a measure of mental rotation ability.

Procedure. Each participant was met separately and was administered the block rotation test. They were then given a brief video introduction to the experiment in which the basics of solving a circuit board were illustrated. The introductory video taught participants the names of each block shape, and illustrated how they could be used to solve a sample circuit board problem.

Participants were then trained on three different circuit board configurations, one in each training medium. All possible orders of the three board configurations and colors were assigned across participants.

Immediately after completing the training for each circuit board, participants were timed on their ability to assemble it in the real world. Participants returned a week later and attempted to solve the same puzzles with no further training. They were also timed on this task.

Results


For each training medium, solution times immediately after training were highly skewed. Subsequent analyses were conducted on a log transformation of these solution times.

The effect of training medium. Differences in the effect of training medium were tested with a 2 (gender) x 3 (medium - Video/VE/Non-interactive VE) repeated measures ANOVA. Neither the (log) solution times immediately after training nor the solution times one week after training were significantly affected by training medium, gender, or their interaction. Figure BM-1 illustrates the different times for the puzzle solutions one week after training in each training medium for men and women. Although performance was fastest with the puzzle that people learned from the video (M = 166 s, SD = 138), it was not significantly different from the interactive VE (M = 221, SD = 195) or non-interactive VE trainers (M = 210, SD = 206) (F(2,42) = 0.52, p = .60).


Figure BM-1



The effect of interactivity. We compared each participant’s performance on the puzzle trained by the interactive VE to that of the yolked participant who viewed this video on his or her non-interactive session. The difference between each participants’ VE solution time and their yolked participant’s non-interactive VE solution time constituted our measure of the effect of interactivity. Mean difference scores both immediately after training (M = -0.04 log seconds, SD = 1.14) and one week later (M = 20.23 s, SD = 326.96) were not significantly different than zero. Figure BM-2 shows the distributions of these scores.


Figure BM-2



Individual differences. Correlations between the mental rotation test and the solution times ranged between -.04 and -.24. None of these correlations were significant. Mental rotation correlated highest with the participant’s gender (point biserial correlation = -.39) however, this was also not significant at conventional levels.

[ return to top ]