I am to compare four groups (two clinical and two controls) in the field of nominal variables in the form 0/1. Is there one test or does the only way to use the Chi-square test for separate pairs of groups? Does anyone have an idea to conduct such an analysis in Statistica software?
It depends on the questions you're asking, but you can either do a global test of independance using chi-square test or suited pairwise comparisons (with appropriate multiplicity corrections).
Impossible to say more without details on the question asked...
Dear Pawel
You can use SPSS software to perform chi-square test of independence,
Please, use the following link, I hope will be useful for you:
https://www.youtube.com/watch?v=lvB0HHoihVA
Good Luck
Hi Pawel
The most popular test is Z-test for the difference between two means of clinical group and the control one. However, to write down the exact formula of such test, I need to know the exact situation of your data and whether the clinical groups represent a particular disease or not.
Regards
Thank you all for your answers! I will probably choose a separate comparison between two clinical groups and two controls because such comparisons may contribute the most to the study I am working on (neural network analysis of the number of hubs and non-hubs in groups of patients with schizophrenia of varying length of disease on the background of changes in this range occurring in healthy subjects differing in age).
Analiza wariancji ANOVA (F) i nieparametryczny test Kruskala-Wallisa (H) +test t-Studenta
@ Pawel: reading the complexity of the question and the underlying data, you may have a huge increased value by consulting a statistician to have better models for the data...
Nevertheless, if you start with separate comparisons, don't forget to handle the multiple comparisons issues with appropriate methods, like Holm's method!
- Badges
- Science topic