SUMMARY.A mathematical task was used among 432 junior school pupils in a search for personality-teaching strategy interactions. Factors of sex, ability, anxiety, extraversion, model and teaching strategy were incorporated into a replicated design. An inductive, learnercentred exploratory strategy and a deductive teachercentred supportive strategy each employed the same mathematical models. There was no evidence of strategy-ability or strategy-extraversion interactions but the strategy-anxiety interaction was significant. Linear functions, relating the retained learning outcome to anxiety score, were used as the basis of decision rules for differential instruction. Pupils at the lower end of the anxiety range may be assigned to the exploratory strategy and those at the upper end of the range to the supportive strategy if advantage is taken of the disordinal nature of the observed interaction. The consistency of a negative linear relationship between outcome and anxiety for the exploratory strategy contrasted with the complex pattern associated with the supportive strategy. In the latter case, linear functions of differing sign afforded predictability for subgroups of introverts and extraverts and for subgroups of low ability and high ability children. When all data were combined for the supportive strategy, however, a non-linear relationship was revealed.INTRODUCTION IF a line of regression relating learner-aptitude to learning outcome is drawn for each treatment under consideration, then non-parallelism w i l l indicate the presence of an aptitude-treatment interaction (ATI). An ordinal interaction is defined as one in which the order of treatments A and B is the same over the entire aptitude range, i.e., one treatment is superior overall (Figure 1). Where the regression line for treatment A crosses the regression line for treatment B within the aptitude range, however, the interaction is said to be disordinal and decision rules for differential instruction may be formulated. Pupils at one end of the aptitude range may be assigned with advantage to one treatment, while those at the other end of the range are assigned to the alternative treatment.Where analysis of variance is employed then mean outcome scores will be examined at blocked high and low levels of aptitude. A significant F-ratio for interaction, with crossing graphs of cell means, was formerly considered suffcient for disordinality (Lindquist, 1953 ; Lubin, 1961), but Bracht (Bracht andGlass, 1968) proposed more stringent criteria, conservative in relation to the risk of false rejection of the ordinality hypothesis. Differences between outcome means at both high and low levels of the aptitude variable were required to be separately significant, as well as opposite in sign. Reaction against the severity of this suggestion (Leith, 1972 ;Cronbach and Snow, 1973) has been based upon the extremely low probability of joint occurrence of two separately significant events. A recent distinction has been made between research set up Dr. Trown is now in the Depa...

The investigation was concerned with possible interactions between teaching strategy and personality differences among children learning mathematics in the early years of secondary school. Part of a text in extensive current classroom use was the basis of learning material structured so that the effect of giving rules before or after examples could be revealed.Throughout the experiment introverts were superior in performance when rules were presented before examples and extraverts were superior when rules were presented after examples. This interaction was significant for all three dependent variables of ' original ' learning, ' retention ' and 'transfer.' When two levels of anxiety and two levels of intelligence were examined separately a similar treatments x extraversion interaction appeared within every analysis, whatever the dependent variable, reaching significance in several instances.

No abstract

To test hypotheses about the optimal place of rules in school learning tasks 124 12-year-old children from a single campus were categorised by ability, sex and two personality traits --extraversion/introversion and general anxiety. The learning task was a program on vectors from which rules were 'abstracted and given either before or after sections of the program containing practice examples.Further evidence for the superiority of rules following practice was obtained. Significant interactions of treatments and extraversion on post-and transfer-tests showed, however, that this occurred because the "rules before" was significantly poorer than the "rules after" condition for extraverts of both above and below average ability. There was no significant difference between the treatments for introverts. Anxiety level differences were not significant, but anxious children were slightly better than non-anxious.