Order the means from smallest to largest. Underscore the pairs of means that are not significantly different. The pairs not underscored are significantly different. This presentation goes back to Duncan, Biometrics, 1955, Multiple range and F tests.

Dear Mark,

You can use Excel spreadsheet for calculations. Then. for background. there is a web site and e-book, all for free.

The idea is that you need to arange your data from smallest to largest means, then draw a line under these that do not differ between themselves, eventually, multiple lines, if you have many of them. Than mark each line with separate letter (a, b, c etc.) and, finally, asign these letters to respective means...

regards,

Gederts

Here is an Excel file, pdf and respective link

You can plot significant differences as follows: A-D or A-F, which is depending on the number of significant treatments.

Best regards

Thanks guys for your assistance, let me try and see

Perhaps this paper could help.

Thanks Emma

I think SPSS program is the the best also i think now version 18 or PSW. also another program STAT or COSTAT, but SPSS so easliy to use

The p-value of the anova test is put in table containing means and standard deviations of different groups. Then in the text we can say for example that group 1 significantly varried with group 2 and group 3 non significantly varried with 2 (p

Hi , I guess you have to design a table ( the row and column depend on the number of TX. some times two tables are required. then each cells which contains a significant value should be marked with an alphabet or star......

I hope I got it right. if so i have attached a file to show an example .

Table 3. Multiple comparisons results for the effect of different media on the flexural strength of two cements.

Cement Panavia BisCem

Media NH LA DW OB HB NH LA DW OB HB

NH - .001* .62 .02* .00* - .00* .00* .00* .00*

LA .001* - .14 1.00 .51 .00* - .06 .00* .01*

DW .62 .14 - .51 .01* .00* .06 - .00* .00*

OB .02* 1.00 .51 - .84 .00* .00* .00* - 1.00

HB .00* .51 .01* .84 - .00* .01 .00* 1.00 -

• Significant at p< .05

Post hoc tests are done when you have not defined the groups earlier and an initial exploration reveals that further comparison between groups more than 2 in number is worthwhile. In this case you may wish to use post hoc ANOVA. Since there are multiple groups it is quite possible that some of the groups will be different from each other by chance. This can be addressed by using Bonferroni test, which adjusts for multiple comparisons. In SPSS follow these steps: Analysis - Compare means –One way ANOVA –

Factor – select the variable for stratification (such as five groups of players, blue, red, orange, green),

Dependent list – select the variable you want the means and compare (such as lung capacity measured by ..).

Contrast – polynomial (to take care of distribution of data)

In post hoc – equal variances assumed – select Bonferroni (this will adjust for multiple comparisons),

In options – click all boxes to get a detailed result, you get the means and the multiple comparisons,

Check descriptives to see the means and post hoc tests, multiple comparisons to see which mean value differ from which one taking into account that multiple comparisons are made.

If you want to adjust the analysis with other factors assess carefully whether you really need to adjust.

draw a table with 8 row and 8 column and write in each cell dif. of means , SD and P value for any 2 goups.

Sorry, I misunderstood the question. In terms of presentation, I prefer to present mean, and SE (or 95% CI for the mean), total n and as notes under the table with superscript letters, p values indicating which mean differed from which one.

Order the means from smallest to largest. Underscore the pairs of means that are not significantly different. The pairs not underscored are significantly different. This presentation goes back to Duncan, Biometrics, 1955, Multiple range and F tests.

In addition to the statistics reported, I suggest prepare your ANOVA table as clearly as possible from title to notes under the table. The table should be understandable without referring to the text. Using superscript letters, notes under the table should include the analysis used and p values. Such as, post hoc ANOVA (Bonferroni adjusted), x compared to y, p=0.001.

I've set mine out like a correlation matrix, which has asterisks denoting significance, but instead of correlations the numbers represent the mean differences between job occupations. And I've done a separate table above this showing means & SDs.

I haven't seen it done elsewhere like this before though so it may not be 'legit.'

I would like to know how to get that superscripts letter over the values which denotes the significant differences between each other... Please suggest..because while we doing we get the data with asterisk only which says the significants but in terms of presentation how we can show it in superscripts letter for the significant differences

I use Microsoft word to present my results in the form of tables, use the superscript option there in the toolbar, and carefully match with printed results obtained from the statistical software, and check again and again. Repeat the analysis and confirm the results obtained from statistical software matches with those presented in the tables.

some journals even ask for F df and P value and sign value apart from usual letters depicting significant difference. My question is how to cite them?