Dear Sunethra,

You can do Paired sample T test for your data.

thanks Mohan. I used it to compared means of different visits in the same group and used two sample t-test to compare between groups. However some query that Chi-Square test is the most appropriate since these are binary data. I like to know the opinion of other researcher or statisticians on this.

Respected Madam,

You can proceed your statistical analysis with DMRT (Duncan's Multiple Range Test).

Good Luck

Thanks Tresina

Dear Sunethra,

Are you sure you're dealing with binary (or categorical) outcomes? If you've used means to present the data, it sounds as if you're dealing with a continuous variable. A simple question to ask yourself is this: Is the outcome measured as a discrete number on a continuous scale (e.g. 37.5 units; 100.5 mls; etc.) or as a categorical outcome (e.g. bleeding vs no bleeding) which is expressed as proportions? If it's the former, then you're essentially comparing means (or medians, for non-parametric data); if it's the latter, then Chi-squared tests are appropriate. Hope that helps! 

Dear Thomas,

thank you very much for the answer. I have recorded bleeding as YES/NO (Categorical) and expressed as proportions/percentages. so the Chi-squared test is the most appropriate to analyze my data set. Thanks a lot.