Hello Sara,

                  using either parametric or non-parametric tests is dependent on the distribution of your data. If your data is normally distributed then a parametric test is appropriate; if not then use a non-parametric test. You can test for distributional assumptions using histograms and/or QQ plots. When you mentioned 62 binary variables am assuming your data is categorical. Your outcome will determine the kind of test to run. So, what is your research question?

Hi, Sara,

The best question to ask yourself when conducting exploratory analysis such as you describe is, what is the scale of measurement for my variables, especially your dependent variable(s)? For example, if the six Likert scale items you mention produce a summative value, then the dependent variable could be treated as interval/ratio. The 62 binary categories can then be evaluated on the DV with a simple t-test.

Alternatively, if the six Likert items are to be tested as DVs individually, you could use the Mann Whitney U Test to compare the 62 binary categories.

If neither of these scenarios fit your need, I'll happy to help further if you send me more information.

Dear Udochisom,

Am using non-parametric because there is no normality which is the case usually for binary question types. My variables has yes or no answers mostly so I used scale and am searching for all the possible tests to perform so I will add all them. Thank you for your help

Dear Peter,

I will try Mann Whitney U to test the binary, and for my scale am testing different things so I was wondering if the nonparametric one- Kolmogorov Smirnov test can be used for all variables whether binary or Likert scale questions and if it is a good test?

Thank you  lot for your help

Dear Peter

You can use Mann-Whitney to test two independent variables have a non-normal distribution, in the case of test three variables or more you can use Kruskal-Wallis test.

Good Luck

Dear khalid,

I tried kruskal_Wallis test and it seemed useful to my data. Thanks for your time

what are you actually testing ?

Sara

You developed a questionnaire and applied it to 100 people. For what purpose? Surely you know it but your question above does not reveal it. Are you exploring or describing data, or testing a number of hypotheses? Are you constructing some scales by summing up the items in the questionnaire? The aim of the study is a guide to the answers.