Immediacy and Learning Meta-Analysis 24
reliability of measurement resulted in the expansion of both the estimates and the variance of the
samples. Post hoc analyses were conducted after removing the most extreme cases as outliers,
and variance was, of course, reduced. However, consistent homogeneity within the samples
remained elusive. This indicates that undetected moderator variables may still be present in these
studies, a matter to be pursued in future research.
Conclusion
Continuing research in this area would probably profit from more focus on actual
cognitive gain or increased cognitive performance associated with teacher immediacy. Attention
to higher-order learning or different types of learning, especially over time, would be helpful in
understanding the relations among types of immediacy and categories of learning outcomes.
Further examination of affective measures and motivation measures as mediators of a path from
immediacy to cognitive or behavioral change would facilitate better understanding of affect in a
more complete learning model. The large number of studies involving affective learning warrant
this next step. This research, of course, should take care to report appropriate, obtained
reliabilities and norms. Before such research is undertaken, however, linearity assumptions
underlying correlational results could be tested. For example, the use of quartiles or higher-order
polynomials would facilitate testing the covariation of learning outcomes with teacher
immediacy across all appropriate levels. Finally, confirmatory experimental research would
allow tests of causality in these research endeavors.
Convention