I am doing an experiment where I measure the blood pressure of a treatment and control group over the course of several weeks. At each measurement session I collect multiple recordings. From what I understand, the best test for me to run is a repeated-measures ANOVA, but I don't know how to account for the multiple BP recordings. If it makes a difference, the amount of recordings I can collect from each animal in a given session varies (mice are tricky!)
The software I am using is GraphPad Prism, if anyone has pointers for how to run this test on that program.
Dear Lauren,
you got it absolutely right. Your experiment is a classic example for a RM ANOVA. You have two groups and several time points.
I see you are using GraphPad which is great to start with as it offers sample data and very good guidance. You may have seen that if you start a new file and choose "grouped" you can select "start with sample data and tutorial", below you select "two-way ANOVA" and "RM matched in rows or stacked". Have a look at the sample data and the tutorial. You will see how the data is organized and how you can analyze. Here you can select general linear model (GLM) and mixed-effects model which will also work if you have missing data. Be aware that you ultimately want to have a look at the time x group interaction effect to see if changes are different between groups. After this, look at the main effects.
Good luck!
Boris
Boris Schmitz Thank you for responding to my question! I'm still not sure how to account for both the repeated measures across days as well as within days.
Right now, I have each column as a different day and each subcolumn is collected that day from a different specimen. Each row is one of the 5 measurements collected for the specimen on a given day. I've included a partial screenshot for clarity.
This seems to be the best way to arrange the data so that repeated measures can be matched , but I am frustrated that there is no way to designate which specimens belong to which group (treatment or control). Any pointers?
I see your problem now. I guess you do different days because your hypothesis is that treatment affects long-term blood pressure? Then multiple measurements during the day can just be seen as repeats at the same time point. Then you can define groups in the two-way ANOVA. If you are also hypothesizing that treatment affects diurnal variation, design is three-way. I also see a zero in the data set. You need to remove. Good luck. Boris
Is there a place to define groups in the two-way ANOVA? I suppose that's what I'm getting stuck on. Thank you so much for your quick reply!
No problem; no use in wasting time on problems already solved. Here is the screenshot from the sample data. Should do for your data, too.
Boris Schmitz If each group is a column, then I suppose each specimen should be a subcolumn. Each row would then be a different measurement day. How could I represent the different measurements collected each day?
In this table, you would put all specimen and the daily measurements as replicates in the subcolumn A1-Ax etc. As I said, it also depends on the hypothesis. If diurnal variation at different days compared between groups is a question you are looking at three-way.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling
- Analyses of Variance
- ANOVA