Scenario: Independant variable = Sugar
DV= weight gain
# subjects 50
random assignment into 5 groups, with varying amounts of sugar from 0 to 100gms.
Groups 1- 4 have unlimited access to IV, group #5 is limited to one "dose".
At the end of the week subjects are weighed.
1. What is the omnibus null hypothesis (HO)?
2. If conducting a pairwise contrast, how many possible pairs of means would be compared to one another?
3. Assuming a Type I error rate of 5% per comparison, what is the total Type I error rate (using Bonferroni's inequality)?
1. The null hypothesis is there is no difference between the means of the population divisions for weight "gains" of the subject. However, I do not see in your study where you are weighing at the beginning of the week. How are the 50 subjects "randomly" selected? Is the height to body mass index range used in pulling the 50? Is this a Healthy Human Subject Population delineated by a protocol for what is a Healthy Human per FDA guidelines?
2. It seems you have five equal groups of ten subjects so (a) is five for the analysis of variance ANOVA. The N is 50. You will calculate the mean for each ...