I have the statistical research data. I have 6 treatments with 3 boxes for each treatment. I am raining on the treatments every week and collecting data. The data is variable within boxes and also with time. For eg: on 20 July, value for seeded treatment was 4500 and after the end of August, the value becomes 2000 with the increase in vegetation. The standard deviation of the treatment for the whole study period is sometimes more than mean due to higher variability. the sample size for the whole study period is close to 350. The data comes out to be normal. out of 6 treatments, 4 treatments have very lower values than other 2 treatments. for eg: 4 treatments have the values in range of 50-500 and other 2 have the values in range of 2200-2500. When I performed Tukey Kramer on all 6 treatments, there is no significance among the 4 treatments. When I apply Tukey Kramer on just the 4 low value treatments, there is some significance among the treatments. How is it possible? What exact analysis or test can I perform on this kind of data? I would be highly obliged if someone has some suggestions or solutions. I need to defend soon and i need to finish this statistical analysis.