Here is my study design:
- there are four soil plots - A,B,C,D
- each plot is sampled for soil moisture at 15 locations
- this soil moisture sampling is done for 10 rain events
I need to compare the soil plots A to D for statistical difference of soil moisture.
First approach - I took an average for each location within the soil plot from all 10 rain events - this left me with 15 averages per soil plot to represent the mean soil moisture at location 1, 2, 3, etc. I ran an ANOVA on this dataset, so comparing plots A,B,C,D, where each plot contains 15 averaged datapoints.
Second approach - I ran an ANOVA using the raw data, so each plot contained 150 data measurements (15 locations within each plot for 10 rain events).
Problem - I get different results using the two approaches - one shows statistical significance and not the other.
Which is the correct approach?