“PSY233 STATISTICS ASSESSMENT 1. The table below gives the data from an experiment similar to that described in problem sheet 1, question 3, but using a different scoring system and with more subjects. So the data are social behaviour assessments of five groups of juvenile delinquents, with group 1 placed on probation without any special treatment. (A high score indicates better social behaviour.) Analyse the data fully, including any appropriate post-hoc test. 2. A random sample of 8 subjects were shown four short films each of which were designed to change their attitude to an important or topical issue. The results are given below, with a large (positive) score indicating a change in attitude in the intended direction (a negative score indicates a change in attitude opposite to that intended). Analyse these data fully and report your conclusions.”

At first glance the treatment condition of Group 4 seems to be the most efficient in promoting the highest levels of better social behaviour amongst its juvenile delinquents.  Alternatively the Group 1 condition seems to be the least effective in this respect.  The assumptions made were homogeneity of variance and ANOVA.  Therefore the null hypothesis can be rejected and the alternative experimental hypothesis is accepted.  A follow-up analysis is now required.

A Tukey post-hoc test gave way to the inter-comparison of each p value.  There appears to be a division of the five group conditions into two subsets.  The first subset is composed of the group 1, 3, 5 and 2 conditions, whereas the second subset is composed of the group 3, 5, 2 and 4 conditions.  This shows that groups 1, 3, 5 and 2 all have significance reduction, but no significant difference when compared to group 4; likewise …(short extract)