What is the statistical tool to compare sample means reference to specific sample mean?
How to use SPSS / Minitab S/W for this purpose?
e.g. I want to compare the means of these samples with "Mu" value.
You can perform Dunnet's test. For this, create a categorical group variable representing different mean groups. Related to ANOVA menu, you could find Dunnet's post hoc test option in both programs I suppose. Choose the reference group as your Mu group. SPSS only selects first or last group of your categorical variable as reference group. So when creating your categorical group variable, code Mu group as first or last. After that you can get the results of Dunnet's test. I have not access to Minitab. But I believe it should be similar.
Akila. Perhaps if you divide each sample mean Ui / U you obtain the fractions of Mu and you may compare their distributions and graphs. Some fractions values will be bigger than one, others smaller than one. I ignore what distribution model you prefer to use in the analysis: you decide it. Your result´s certainty will depend on your main sample and your main method, premises and applied models. Best wishes.
Thank you Dr. Mehmet and Dr. Emilio.
Dr. Emilio
How to compare fractions that have equal denominators.
If present the mean of all the genotypes as a fraction of the mean of "Mu" geneotype. so "Mu" will always be 1. To explain this result, is there any statistical method?
I asked this few weeks ago but didn't receive satisfactory answer.
https://www.researchgate.net/post/How_to_compare_fractions_that_have_equal_denominators .
Akila. It would be similar to method in trigonometry: you use real value and units of U when you know theoretical values of model (sin, cos, tan, etc.). In statistics you trust your media U is close to real value, and asume your distribution model (in medias) as representative, so you decide to apply them together. Modeling in statistics is more complex and have risks and uncertainties, but it is possible. Otherwise, induction or "regression" -as europeans called it- would be impossible in science.
You measure and work genotypes, so I guess that there are families of genotypes that depend on measured size, either in the units you use, or as fractions of U obtained from sample. So you must be clear about it (I am no expert in that field). OK, success for you.
Hallo and Dear Akila
you have to use SPSS by comparative test model, may use the mean or one sample t test .. following steps: Select Analyze Compare Means + + means or the One Sample t test. see attached my data.
reference to a cultivar or to an environment does not imply a relative scaling. The mean separation is run by clustering and several references are described in a recent report, http://www.sciencedirect.com/science/article/pii/S2214514115001051
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS