Results are presented separately for items involving the teaching of commonly shared
values and less commonly shared values. Evaluations and justifications were analyzed using
analyses of variance (ANOVAs) on major vaxiables of interest. Proportionate usage of
justifications was calculated and corrected for nonnonnality with an arcsin transformation
(Winer. Brown. & Michels. 199 1 ). Only elaborated justifications comprising 10% or more of
responses overall were analyzed. For justifications. where several ANOVAs were performed.
the alpha lrvel was set at .O1 to control for type 1 errors. Post-hoc cornparisons for al1
analyses were performed using the Bonferroni test, with critical values set atp < .O5
Preliminary analyses revealed no significant main effects or interactions of order for al1
analyses and no significant main effects or interactions of gender for the evaluations:
accordingly. these variables were excluded From subsequent analyses. (Gender was significant
in the justification analyses and was thus included as a variable there.)