put BAP, Kn, TDZ and 2-ip in column. number or growth of axilllary shoot in row (as it is the variable u want to see the effect of those PGRs). do one-way ANOVA. then see the mean and compare the P value to be assigned as "a" "b" "c" "ab" "ac" etc, this part u have to do manually by seeing the table in any stat book or u can download from net.

thnx.

Yes, as Santosh says, in the data view, enter one column with the condition (BAP, Kn, etc) and one column with the outcome (shoots per node or shoot length). Then you can do an ANOVA using the ANOVA command from the drop-down menu. One row for each plant.

@ madam Priya Kalra, thnx. bt the solution can be best if it will be calculated manually for "a" "b" "c" "ab" "ac" etc,

If you have little knowledge about data entry in SPSS that it would be good for you to first copy a set of data from MS Excel and see how the data appears. That would give you an idea as to what is data view and what is variable view.

it is easy to use SPSS soft ware version 16 however first you have to arrange the data in a column in the working sheet of the spss software then you can do ANNOVA one way or two way by following the commands from the drop-down menu. How ever if you are beginner then you may use the software biostat or irristat available freely in internet. These are easier than SPSS.

You do not want to do a one-way ANOVA if there are four treatment factors. Since you are talking about a, b , ab, c, abc etc. (which are interactions between the treatment factors) I am assuming you want to test main effects and interactions; it will be a four-way ANOVA and you want to include higher-order interactions. You must have five columns in the DATA view, one each for BAP, Kn, TDZ, 2-ip and Number of Shoots. You enter the data one line per observation (replicate). In the VARIABLE view you give them names (BAP, Kn etc.).

To do the analysis you go to ANALYZE - GENERAL LINEAR MODEL - UNIVARIATE. Move Number of Shoots (your response variable) into the DEPENDENT VARIABLE. Now you have to distinguish between Fixed factors and Covariates (forget about random factors for now). A fixed factor is a treatment with a discrete number of values (such as dosage) and a Covariate is a measured value (such a height of plant). You must decide which of BAP, Kn, TDZ and 2-ip are factors and which are covariates (I am not a geneticist so do not know what these quantities are) and move each into the appropriate block. Once you have done that, you click on MODEL and you are given the option of fitting the full model (including all interactions) or to make your own choices of main effects and interactions. Under POST HOC you have a choice of several multiple comparison methods (including DUNCAN) but these Post Hoc tests are done only for Factors (not for Covariates).

I hope this will be helpful.

IN a one way analysis with DMRT you obtain a table like this:

VAR00001

Duncan

Subset for alpha = 0.05

VAR00002 N 1 2 3

1,00 2 2,0000

2,00 2 3,5000 3,5000

3,00 2 5,5000

4,00 2 8,5000

Sig. ,184 ,099 1,000

Means for groups in homogeneous subsets are displayed.

a. Uses Harmonic Mean Sample Size = 2.000.

It is easy to type the "letters" I.E col 1 is a, col 2 is b, col 3 is c. If a means is in 2 cols this means have 2 or more letters.

SO IN thE PREVIOUS table the letter are the following (using a for the smallest values) :

Subset for alpha = 0.05

VAR00002 N 1 2 3

1,00 2 2,0000 a

2,00 2 3,5000 3,5000 ab

3,00 2 5,5000 b

4,00 2 8,5000 c

(it is not your question, but why DMRT ? Is is not a good test, because do not control the type I error rate ) . There are several papers about this. UNlukely DMRT is still widespread in agronomic and horticultural research.

I totally agree with Marco. please find the attached documents as examples (Raw data first, followed by the overall analysis and how I put together the output in a graphical representation)

@Kher, The raw data,SPSS version of data and analysed results documents that I uploaded with my publication are not true values (they are just examples guiding you on how you can enter and analyse your results) The mean values, std dev and std error under descriptive section can be assigned different letters according to subsets that they belong to (Like how Marco described above) I will not give true data for what you requested as those results are not public. Thanks and wish you well.

The SPSS,is programmed to have Data View and Variable view, so i see the kind of data as qualitative kind of data, which requires several outcomes in the data, e.g, Yes, No, severely, less severe, e.t c, that means he will have to do Non-parametric Test to conduct Yetes, correction, or chi's square and others, as the case may require, many more outcomes and you can test for significance level depending on the statistical relevance of your data sample size, here the issues of sample size arises.