Dear Dr.Khalil

With my kind of understanding of your question that a particular question has multiple options and how to code them in SPSS. I think it is easy before you think about any analysis.

The example has 13 options, each of this response goes to a single column, and it can be entered as short sleeve : yes or no;

long sleeve: yes or no

shorts: yes or no and so on

At the end of it you can see how many are saying yes (or wearing) short sleeves, long sleves, shorts and so on.

good luck with your coding

thanking you

I do not know what you means. If you are coding factors, you should check the level of your factors. Example if you have two level: just code the extreme high and low. if you have more than two level: just code the extreme them take the distance between them and try to code at equal distance depending on the level. This method is very good in order to optimize your responses. It does not matter if you have several factors and you need to code all of them you can do that at different level and at full or half factorials or 1/4 factorials..

Your responses will be checked depending on the half percentage probability and the ANOVA. No matter how many responses you have: some will be significant and some not.

Hi, Khalil,

The example you have provided is a series of item responses on a nominal scale of measurement without rank order. It\\Each individual article of clothing receives a ye/no response and should be coded that way in SPSS. It might be possible to create some summative values from these items based simply on the number of items worn, but that is a question only your data can answer. Reliability tests in SPSS may show you those items that are highly correlated and suitable for grouping as will an exploratory factor analysis, although I would have to check on the suitability of EFA with nominal data.

Grouping in SPSS is easy once you understand the response

Dear Dr.Khalil

With my kind of understanding of your question that a particular question has multiple options and how to code them in SPSS. I think it is easy before you think about any analysis.

The example has 13 options, each of this response goes to a single column, and it can be entered as short sleeve : yes or no;

long sleeve: yes or no

shorts: yes or no and so on

At the end of it you can see how many are saying yes (or wearing) short sleeves, long sleves, shorts and so on.

good luck with your coding

thanking you

Thanks everyone for your help.

I also think that using DOE from stat ease is better than SPSS

Dear Khalil,

Following Ravi's coding advice I would recommend analysis using Cochran's Q. This is one of the standard nonparametric tests for repeated measures but it wor with yes/no answers.

I worked several factors and responses using DOE and SPSS and found out that DOE is much better

check some of the articles in tribology transactions and wear and see how many variables were used to optimize lubrication conditions and wear.

Answer provided by Rita Pitani about coding each answer in form of yes / no can provide answer to each response about how many answered yes and no for each items. in SPSS two type of analysis of multiple response is available - frequency and crosstabs which provides percentage to total respondents and percentage to total response. No statistics. Hope this will be useful

Multiple responses depend on the half normal percent probability and the ANOVA

I support Ravi Pitani response. Apply it.

I am sure it should answer your question.

You should look for the most probable cause in the interactions domain and the overall model statistics

Ravi Pitani gave a good response. do not forget to add a value for missing entries too

You should apply the law of multiple interactions when you enter your data

Dear Sir,

As far as my understanding, first one need to understand what is your objective and how these variables will serve the purpose.

Further, DOE can be of better help if you are clear abt your objectives as you can find out interaction between the variables in a easy manner rather than a simple coding.

Look for options to optimize your model

I have always conducted such analyses using Ravi's recommendation. It's ok by me.