topic Re: Question in regards to Tukey's Multiple Comparisons Test in Discussions

Question in regards to Tukey's Multiple Comparisons Test

ana — Wed, 22 Nov 2017 11:09:45 GMT

Hi there,

Would anyone be able to tell me how to go about running a Tukey's multiple comparisons test on my data? I've collected soils samples from 3 farming systems (n=3) on day 0, 14, and 28. I'm measured the cfu. I'm trying determine whether or not there's a signifacant difference between the 3 farming systems on each collection date. I know there's no sig. dif. on day zero and I'm pretty sure there's a sig. dif. between the farming systems on day 14 and 28, I'm just not sure how to go about running the stat. analysis to prove that. I've inserted a photo of a graph and my data. Any help will be greatly appreciated!

Ana (my apologies if I posted this message in the wrong place)

Re: Question in regards to Tukey's Multiple Comparisons Test

shoffmeister — Wed, 22 Nov 2017 12:59:55 GMT

You can use a two-way-ANOVA the following way:

1. Main Menu => Analyze

2. Fit Model

Fit Model Dialog (I like to use "Minimal Report")

3. In the fit-model dialog: Y is CFU

4. Select Farming System and Day => Macros => Factorial to Degree

5. Run

Then go to the main Hotspot (red triangle) next to Response CFU and select Estimates => Multiple Comparisons.

Select the interaction term and "All Pairwise Comparisons - Tukey HSD". That will give you some tables and graphs to figure oput which groups are significantly different.

Hope that helps,

Sebastian

Re: Question in regards to Tukey's Multiple Comparisons Test

ana — Thu, 23 Nov 2017 10:34:58 GMT

Thank you SO much Sebastian! I really appreciate you taking the time to work out the steps. Ana

Re: Question in regards to Tukey's Multiple Comparisons Test

Leycoste — Thu, 16 May 2019 09:51:26 GMT

Hello and thank you for this explanation.

I had the same problem (ANOVA + Tukey, with weights in mices depending on the treatment).

However, can you help me to interpret the significant differences between the groups for each time?

what data should I look at for each week?

Thank you.

To be clearer, I would like to know the p-value between each group for each week

PS: I had no choice but to select the treatment on multiple comparisons

[EDIT] : Problem resolved. My variables (week) was not in "nominal".