Thank you for explaining all that and helping me with my data. To answer your questions..
You have each treatment "run" multiple times. Are these multiple animals? Or the same animal "run" over different time periods?
I placed one animal into an arena for 10mins and tracked its locomotion. For example, animal1 traveled 196mm during the first minute and between min1 and min2 animal1 traveled 1595mm, ect. So each data point is a sum of the milimeters traveled during that 1min time frame. For the NCS treatment group I had 24 different animals that received the same treatment. For the CS1 treatment group I had 24 different animals that received that treatment. For the CS6 treatment group I had 25 different animals. So the treatment combination was the same and have multiple data points so these are repeats, as you said.
It looks like you averaged these? Why?
I’m not that interested in the individuals. I’m most interested in how the treatment affects locomotion. So that is why I averaged them so I could compare across treatment groups. These seems to be what others in the field are publishing in the literature. (For example: Figure 4 in Lind et al 2005 Validation of a digital video tracking system for recording pig locomotor behaviour)
Did you look at the within treatment variation before you averaged them?
I know there is quite a bit of variation within a treatment group, which is why I have pretty large error bars. I also know that my data is not normally distributed which is why I considered doing a Freedman’s Rank test. However, others in the literature did a repeated measures ANOVA or at least some sort of ANOVA that they didn’t specify (For example: Figure 2B in Martin 2004 A portrait of locomotor behaviour in Drosophila determined by a video-tracking paradigm) I figured that since my sample size was over 20 it would be okay to do a parametric test instead of a non-parametric test (Freedman’s Rank). Do you agree?
And always look for unusual data points before summarizing.
I know that during some minutes a given animal might not move at all (0mm traveled). This makes for large max/mins and large error bars but since it occurred in all treatment groups, I figured it was okay to average the data for each treatment group for each minute.
Average the data across 10min: I have a different graph that captures this idea (attached box and whisker plot)
Standard Deviation: I don’t report this number but I did show the SEM in my error bars
Min & Max: I show this in my box and whisker plot which is separate from the line graph I’m working on now
Slope of the line: I think this would be interesting to look into since it appears that across all groups animals tend to move less as time goes on. I would probably have to find the best fit line.
Thank you for showing me how to set up my data in the stacked document you sent me. That was very helpful.
“JMP offers multiple methods to analyze repeated measures: a multivariate repeated-measures approach, a univariate split-plot approach, and an additional capability through JMP® Pro to perform such an analysis with the Mixed Model personality within the Fit Model platform.” I’m not sure which of these methods to use. It looks like the Full-Factorial-Repeated Measures ANOVA add in might be helpful if you know what you are doing. I plan to learn how to use it since I have 35 more graphs very similar to this one that I need to do stats for.
Thank you so much for all your help. I really appreciate it.
