Information Technology Reference
In-Depth Information
asked a deep question about the concept, second the virtual tutor presented a content
statement that elaborated the meaning and use of the concept, and third the virtual
tutee provided a “self-explanation” of the concept. The three-event sequence involv-
ing the next concept was then presented. A total of 50 of these three event-sequences
was presented. In an
explanation
condition the first event was omitted from each
sequence. In a
question
condition the third event was omitted from each sequence.
Finally, in a
monolog
condition both the first and third events in the sequence were
omitted. That is, only the virtual tutor's content statements containing the concepts
were presented in this latter condition. The two agents were located on each stu-
dent's monitor throughout in all four conditions, and a series of images appropriate
to the content were located between the two agents.
Analyses of the pretest scores yielded no differences among the four experi-
mental conditions at either the high or low knowledge level. Analysis of pretest-to-
posttest change scores yielded a significant interaction between knowledge level and
experimental condition. In the question + explanation condition the low-knowledge
learners (31% gain) significantly outperformed those in the high-knowledge condi-
tion 7% gain). The low-knowledge learners in the question + explanation condition
outperformed those with low knowledge in the question condition (21% gain),
and the difference approached significance when they were contrasted with the
monolog condition (23% gain). It seems, consistent with some previous findings
(McNamara & Kintsch, 1996; McNamara & Shapiro, 2005), that more knowl-
edgeable learners were unable to take advantage of that knowledge in a condition
(i.e., question + explanation) that enabled greater learning gains among those with
lower knowledge. Among those with high knowledge, the question condition
(21% gain) significantly outperformed the question + explanation condition, and
performance in the latter condition was also exceeded by the two remaining con-
ditions (i.e., monolog
=
19%, explanation
=
14%), but the difference was not
significant.
A second preliminary study of the vicarious explanation effect was conducted
in high-school physics classes. Three conditions were used: question + explanation,
question, and monolog. A total of 50 three event, two event, and single event
sequences were used in the respective conditions. These were presented via lap-
tops during the first part (about 18 min) of each of seven consecutive daily physics
classes. The vicarious presentations were intended to provide conceptual under-
standing of the content presented by the regular classroom physics teacher during
the remainder of each session. The classroom teacher provided equations and appli-
cations that gave quantitative expression to the conceptual content presented earlier
in each daily computerized presentation.
Results revealed that learners in the question + explanation condition signifi-
cantly outperformed those in both the monolog and question conditions, which
did not differ from each other. Learners in the question condition did not outper-
form those in the monolog condition, which failed to support some earlier findings
(Craig et al., 2000, 2006; Driscoll et al. 2003; Gholson & Craig, 2006). It should
be pointed out, though, that more than 50 questions were generally presented in
each session when the deep-questions effect was obtained. For example, Craig et al.
