Statistical Significance with Friedman's Test for Data Analysis"

[jophysics]

Dear all,
I need to perform a statistical test to give statistical significance to the results of my experiment. More specifically:
I have 3 groups of subjects (each group having 14 subjects coupled in 7 pairs) trying 7 different treatments. Each pair in each group performed 7 trials (one for each treatment). The order of presentation of the 7 trials was randomized in each group using a Latin square approach, but the order was the same for each group.

I would like to evaluate which of the treatments is the best and if the differences among these treatments are significant.

My idea was to use the Friedman's test (the normality assumption is not verified), but I am a little bit confused on how I can use it... I checked also the R documentation but I have some doubt yet.

Someone can help me?

thank you

netrunner

 

[Jhero]

The Friedman test would be appropriate for your study. You can use the R software to calculate the Friedman test statistic and p-value, and then interpret the results. You should also check the assumptions of the test to make sure that you are using it correctly. Generally, the assumptions are that the data are ordinal (not continuous) and that the subjects are randomly sampled from a population. Additionally, the observations should be independent and the sample size should be sufficiently large so that the distribution of the test statistic is approximately chi-squared. The output of the R function will include the test statistic, the corresponding p-value, and a summary of the results. Once you have the output, you can interpret the results and make conclusions about which treatments are best and whether the differences between them are significant.