I disagree that the type of analysis must be driven by the type of data. The analysis must be driven by the scientific model. 

We cannot answer your question because your question is not stated fully. It should read "I want to see the relationship between each of five factors and …" 

And here you must think about it. Do you want to see how well the factors differentiate between normal and disorder, or within disorders? And are some potentially causative while others are confounders?

Plan your analysis as questions about association and causation. For example "Does grip strength predict falling independent of the person's age?"

Then you can look at the data and decide on the most appropriate statistical procedure, given the way in which each variable has been measured.

If I understand well, you want to know, for each of your influencing factors, the fraction of healthy and unhealty patients. Am I correct?

There are indeed plenty of statistical methods for that, e.g. cross-correlation matrices. If you describe with more details how your available data is structured I may think about more specific solutions.

Let me know

Giuseppe

Consult a statistician!

I think you want to analyse the infulence of different factors in your experment (like enviroment vs genetic factor). Please I recommend you to read stat book (Gomez and Gomez) on 241-pp and follow the steps.

Most likely a case control study, STATA is very good software for the analysis. 

http://www.stata.com/meeting/australia11/abstracts/ausnz11_muller.pdf  (please see link-may be helpful)

The type of data must determinate the type of statistical analysis that you perform. Consult Clifford´s Biostatistics for the Health Science, is the best reference in the discipline. http://www.amazon.es/Biostatistics-Health-Sciences-Clifford-Blair/dp/0131176609

I disagree that the type of analysis must be driven by the type of data. The analysis must be driven by the scientific model. 

We cannot answer your question because your question is not stated fully. It should read "I want to see the relationship between each of five factors and …" 

And here you must think about it. Do you want to see how well the factors differentiate between normal and disorder, or within disorders? And are some potentially causative while others are confounders?

Plan your analysis as questions about association and causation. For example "Does grip strength predict falling independent of the person's age?"

Then you can look at the data and decide on the most appropriate statistical procedure, given the way in which each variable has been measured.

From the information given you would want to do a multiple variate linear regression in order to determine 

With the information given, it would be reckless to suggest a statistical method... What was the scientific basis for the experiment? What kind of assumptions were made? How were the data collected? Are the factors quantitative, qualitative, discrete, continuous? How many responses are there? Are they quantitative, qualitative, discrete, or continuous? ...

Dear Avinash,

In addition to Noel's comment, it would be helpful to attach a sample of you data and description of each variable (e.g. discrete, continuous, categorical, etc.). With this information, we may be able to suggest a few methods that you may bring forward to a statistician. In terms of statistical software, SPSS may be limited in terms of modelling in comparison to other software (e.g. STATA, SAS. and R).

Also in addition to a clearer description of the study objective/question and hypothesis, it would be helpful to know your sample size, are there any repeated measure (e.g. before and after an intervention on same individual), and correlation between observations (e.g. clustering effect). This type of information can help inform modelling technique.   

There are a number of considerations mentioned above that I did not repeat, so please do consider the previous comments.

hi Avinash

any multivariate statistical analysis will do however it depend how deep or how details you want to interpret the outcomes of the analysis ( how strong is the relationship or you want to rank the dependent variables etc).

For examples SVD, PCA, ICR or the simple analysis is correlation coefficient or coefficient of determination. or you can use ANOVA analysis.

Let's wait for Avinash to post here a more informative description of his experiment. Else, we are guessing what needs to be done. 

Dear Avinash,

Consider using the Binary Logistic Regression model for your data. This can be conducted easily using SPSS where your dependent variable takes two values 0 for the control group and 1 for the patient group and you have 5 different factors as independent variables. You may then check the influence of the 5 different factors and may be their odds ratioes too. 

@Mahmoud : Thank you for your suggestion. I am considering your way to deal with the problem. 

@Raid: Thanks. I m still not yet done with the data. When I finish it, I will again seek for your all help.

@Ronan:  I want to see the how well the factors differentiate between well and disorder.