I am studying 5 different groups of seeds (1 control group and 4 scarification groups, different intensity) oer time, ie repeat measurements (7 measures over one week) (treatment = independent variables). One dependent variable, ie germination of seeds, which can be expressed in yes/no or in days until 50% of seeds show germination. Which is the best statistical test to use? Repeat measurement ANOVA? Other tests?
Thanks
I agree to Alexander, just want to add that such models are also known as "hierarchical models" or "mixed-effects models".
Book recommendation: Gelman & Hill: Data Analysis using regression and multilevel/hierarchical models. Cambridge University Press.
If your responses (dv) come from nominal values (yes/no) you can try a logistc regression model for repeated observations.
https://www.cscu.cornell.edu/news/statnews/stnews22.pdf
hi !
i have similiar problem -
What is the best statistical method for repeat measurement over time with different groups? but with covariates like cell count which i need to adjust.
i have pre and post methylation data . group 1 - pre and post , group 2 - pre and post.
i have cell count of pre and post for each group as well.
i want to adjust for the cell count before i compare pre and post .
finally objective is to see which group 1 or 2 show difference in methylation across pre and post data. Pre and post are two time points.
any help would be really appreciated.
Thank you
smeeta
That does not require mixed models. You have a "simple" two-factorial model with the factors "time" (levels: pre/post) and the factor "group" (levels: #1/#2). The value you are interested in is the interaction of these two factors.
Thank you for the reply.However I need to adjust for covariates that are blood count data. I have blood count for each individual pre and post as well.Where can i include these covariates in the two factorial model.
Maybe you can perform a dependent T test for each group and see in wich group exist statistical differences between pre and post conditions and choose them. Then, you can compare the difference of blood count (post - pre or pre - post, it depends on you) between those groups. I think this could be useful if all the individuals started with a similar blood count
Thank you all for your inputs. I think i will write my problem again.
objective : to understand as which model to use for the analysis :
The study :
1. group 1
> meth at probe1 +cell count data at time point 1 pre
> meth at probe 1+cell count data at time point 2 post
2. group 2
> meth at probe 1+cell count data at time point 1 pre
> meth at probe 1+cell count data at time point 2 post
I analysed methylation difference between pre and post for group 1 and group 2 separately.
group 1 - compared pre vs post
group 2 - compared pre vs post
we did paired ttest.
we observe significant difference in the pre/post data in group 2, where a drug was administered.
However now i want to adjust the methylation data at probe 1 for blood count (lymphocytes,monocytes and granulocytes).
we want to see if adjusting for cell count we still see significant difference in group 2 for the probe 1.
Please could any one tell me as what analysis or regression model should i use.
any help would be greatly appreciated.
regards
smeeta
Just to make sure...you have different initial values for both groups and you want to adjust the data (make blood count similar in both groups) for a better comparison?
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling
- Analyses of Variance
- ANOVA