Two groups ( controls and test) of animals were treated with a particular drug (dose=x g/kg body weight) over a time period of 5 weeks. The blood level of drug is measured at time points 1, 2, 3, 4 and 5 weeks, after sacrificing the animal. The drug level in the each one in test group is expressed as a fold change over control. So now i have 5 groups (1,2,3,4 & 5 weeks) of test animals with blood drug levels (expressed in fold change over control). How can i statistically analyse these five groups? Are these five groups are dependent on each other or not ?
No, they are not depending on each other (at least by no reson I could imagine from your explanations). The 5 values recorded for one animal over the time course are dependent, because thay all share a common source of variance: the animal.
I would say... perhaps, it depends. Just to clarify, you do have 5 groups of animals followed 5 weeks each group? How many animals in each group? Let's assume 10. Why do you need groups? If there's something that can potentially change between groups, such that the 10 animals in a group share some otherwise unmodelled variability? Group might be then better treated as a random effect if you expect fluctuations between groups? An example might be groups of fish in an aquarium, for which subtle uncontrolled differences across k different aquariums might be accounted for using aquarium as a random effects say.
@Tiago Marques and @Jochen Willhelm: Actually there are only 2 groups (controls and tests). Sorry, that i mentioned as five by mistake. Now i have corrected it.
Why do not you do an analysis using repeated measures theory?
You have 2 groups (test and control) and 5 repeated measures (weeks) for all study animals.
- Badges
- Science topic
- Similar topics
- Statistics
- Time Series