Discussion Solution — Mollusks and Litter Decomposition

This experiment produced six independent sets of observations, so the appropriate analysis was by ANOVA or equivalent. The assumptions of equal variances and Normality both are somewhat shaky but I think are acceptable, given the advantages of ANOVA for the follow-up analyses needed to answer the questions. The first question could be answered by either of two methods, depending on exactly how you chose to interpret the question. First, a pre-planned contrast of the control to the average of the mollusk treatments would answer the question in terms of the average mollusk. This contrast indicates that decomposition was greater on average in the mollusk treatments than the control (P = 0.0038). Alternatively, Dunnett’s multiple comparisons could be used to compare each mollusk treatment to the control; these show (at overall α=0.05) that only the L. maximus and O. alliarius treatments were significantly different from the control. The second set of questions could be answered by unplanned multiple comparisons following the ANOVA, or by ANOVA or equivalent with the control treatment excluded; any of these produces the conclusion that there are no statistically significant differences among the species’ effects. The conclusions, thus, are that yes, presence of mollusks increased litter decomposition, but the effects of the species were not significantly different from each other.