Mixed Model Nested Anova or Two sample test?

#1
Hello Forum Participants,

My dissertation research evaluates the effect of a hormone on the epithelium. My data was obtained from a clinical trial with a placebo group and 4 different treatment groups. For each group I have pre and post treatment samples but an unequal number of data points.
I want to compare the pre and post treatment among individuals. The post treatment samples were taken from three different sites of the foreskin, while the pretreatment sample was taken from only one area. Should I use mixed model nested anova for the comparison?
I also want to compare the placebo group against all of the four treatment groups. Initially I thought that I could use the two sample unpaired, unequal variance test. After reading through information on choosing which statistical test to use, I do not think it is the appropriate test.
Any help would be greatly appreciated. Thanks in advance!
 

CowboyBear

Super Moderator
#2
Repeated measures ANOVA might be the way to go, though it depends on the nature of your measurements. You haven't really described your dependent variable... what characteristic of the epithelium are you measuring?
 

CowboyBear

Super Moderator
#4
RM Anova could work then. It does assume normality of the DV within each cell though (5 groups x 2 time points = 10 cells). Variances are also assumed to be homogenous across the cells.
 
#5
pretreatment treatment are
Site one
Site two
Site three

Are four different levels (Xs)

For thickness of the skin layer (y) dependent

In repeated measure anova the independent variable (X) has categories called level

Subject to experiment (say) Treatment location site
Site one Site two Site three pretreatment treatment
S1 (SLTh)1 (SLTh)1 (SLTh)1 (SLTh)1
S2 (SLTh)2 (SLTh)2 (SLTh)2 (SLTh)2
S3 (SLTh)3 (SLTh)3 (SLTh)3 (SLTh)3
S4 (SLTh)4 (SLTh)4 (SLTh)4 (SLTh)4
‘ ‘ ‘ ‘ ‘
‘ ‘ ‘ ‘ ‘
‘ ‘ ‘ ‘ ‘
Sn (SLTh)n (SLTh)n (SLTh)n (SLTh)n

* skin layer thickness (SLTh)


The repeated measures ANOVA tests for whether there are any differences between related population means. The null hypothesis (H0) states that the means are equal:
H0: µ1 = µ2 = µ3 = µ4… = µk
where µ = population mean &
k = number of related groups.
The alternative hypothesis (HA) states that the related population means are not equal (at least one mean is different to another mean):
HA: at least two means are significantly different

In case if there is significant difference in repeated measure anova then we understand that there is a change through medicine or treatment. After this to know which section showed what change we need to run post-hoc tests that can highlight exactly where these differences occur.