Dear list,

I have a sample of 500 respondents, and they are from 19 industries(variable: IN), there is a variable called financial activity(FA). Now I want to collapse 19 industries in to fewer categories based on respondent's answer on FA, i.e., retail, trade, and restaurant industries might share silmiar opinions on their financial activities, so these 3 industries are in one category. Which analysis shall I use then?

[此贴子已经被作者于2005-10-20 12:40:38编辑过]

So FA is the dependent variable and IN is the indenpent variable?

Or, FA is DV, and I can have 19 industries coded as 19 dummy variables, then treat 19 dummy variables as IV?

Can you please explain your logic using GLM, or how will you handle data by using GLM?

actually, it is not a question of regression, why you bother using dependent or independent?

In fact, this is the recoding procedure. You can simply create a new varible, say new_industry, based on ur respondenses, even in excel you can do it.

Which sofeware you are using?

Thank you for the hints!

Clustering: today I calculated mean FA for each industry and treat 19 industries as 19 cases. So I did clustering 19 industries based on mean FA. But my boss is concerned about small variances of mean FA by this technique.

Xiaoyi mm, I don't quite understand what you mean, can you explain more? What is the new variable for?