I want to compare how does a dependent variable (total flow) changes with a covariate (total rain) between two groups (stream A & B).
Usually, i would go with ANCOVA. But these groups have the same number of observations. Each observation in A correspond to an observation in B, because they were generated from the same rain event (total rain). Do i:
If method 2 is better, i have tried and got some negative values from A-B. How do i go about it?
I attached a simplified picture of how my data looks like. But the real data is large (n = 300).
Thank you!
I am irritated by your design, why do two independent groups have the same value for total_rain?
Hi Anand
Simply do two regression analysis, one with flow-a as dependent and total rain as independent, another with flow-b as dependent and total rain as independent. Then compare the two regression coefficients (beta). If they were significantly different then the flow-a responses differently to rain compared to flow-b.
Hope it helps.
Ehsan Khedive Why two regression analyses? This could simply be done in one analysis, if needed, and you could account for the interdependencies of the variables. But still, the design is not clear to me.
Ehsan Khedive Thank you. I would do it the way you suggested. That is the immediate instinctive way. But am also aware that there are other methods used by other researchers. Just want to see if anyone prefers another method.
Rainer Duesing Sorry for the irritation. The reason why both river flow A and B have the same covariate (rain) is because those flows are in adjacent rivers generated by the same rain event in the area.
Rainer Duesing I agree. Doing two regression analysis can be done by one analysis too. I wanted to make clear for him what to do. Besides, @Anand you regression is non-linear.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Regression