I have run Post Hoc test Duncan through SPSS. How it can be explain ? Whether significance level also matters? There are 3 category (small, medium and large farmers) and all category results different subset with 1.67, 2.78 and 4.19 value and significant level is 1.000 in all three cases. Size of the farmers has been compare with financial position (5 point scale very week to very sound).
I am not sure I completely understand your data or model, but let me see if I can provide a response based on what I think you mean (please correct me if I am wrong):
You have run an ANOVA (or regression) with 3 categories (farmers) and an outcome which is on an ordinal scale (1-5). You then run a post-hoc analysis to determine if there was a statistically significant effect between any of the groups, using the Duncan adjustment for multiple testing. You then found that there was no effect between any groups (p=1.0)?
If this is what you did and these are your results, then I am assuming one or more of the following: your sample sizes are small and/or there is substantial heterogeneity in the scores within each group.
Can you provide us with sample size, mean/sd min/max values for each group? This may help us point to the problem. You should also visually inspect the distributions among the 3 groups.
I hope this helps
Ariel
I agree with Dr. Ariel, and I add that there is something not accurate in your post, when you said the scores for the 3 groups were 1.67, 2.78 and 4.19; this means there is variation in the result therefore, it cant have significant value of 1.000, you may here need to check the significance level correctly.
looking for more details on your results and your data.
however, in post hoc, when there are 3 groups, the test gives you the first group in comparison to the other 2 groups, and after that repeat the process with group 2 to be the first, then group three to be first or the reference group. so there will be repetition which you have to ignore.
Dear Dr Ariel Linden, could you please list the assumptions of Ducan test, if any.
Is there any limit on the number of treatments to be compared using this test..
- Badges
- Science topic
- Similar topics
- Biological Science
- Kinesiology
- Running