4. Discussion
One of the clearest results of the many studies that have examined conditional reasoning is the strong effect of content on the inferences that both children and adults make to what are formally identical premises (Cummins et al., 1991; Cummins, 1995; Markovits & Vachon, 1990; Thompson, 1995). A key component of this variation is the effect of the relative numbers of alternative antecedents that are suggested by premises on reasoning with the AC and DA forms, an effect that is replicated in the present study. Importantly, there is also evidence that suggests the existence of four types of premise, which might constitute a developmental hierarchy. These are (1) category-based premises, (2) causal premises, (3) contrary to fact premises and (4) abstract premises. Previous studies allow the conclusion that reasoning with contrary-to-fact and abstract premises represent a qualitative transition compared to reasoning with familiar premises (Markovits & Lortie Forgues, 2011; Markovits, 2014). Within this globally more abstract level, these results also show that reasoning with abstract premises is more difficult, and of a higher level of abstraction, than reasoning with contrary-to-fact premises. The results of the present study allow some conclusions that complement this analysis. They show both a presence of a clear developmental increase in logical reasoning between Grade 3 (8-years of age) and Grade 4 (9-years of age) accompanied by a continued strong influence of premise content throughout the entire age range studied here (see Fig. 1). Levels of logical reasoning are clearly greater for categorical reasoning than for reasoning with causal premises. In addition, within each class of premise content, higher levels of logical reasoning are found with premises that have relatively many alternatives than with premises that have relatively few alternatives, consistent with many previous results.