Dear cheung,,,first of all let me say one thing..before constructing the questionnaire you shold think the analysis part of the response data.the resercher only will be knowing the type of responses from the respondent.It is better to include such type of respnse in he questionnire for future anlysis of the data.here your problem is to code the type of family according to the response given by the respondent.see the response may be multile .I can suggest one way ..ie you go through the questionnaire respons of this particular question and fow frequent the respondent given a particular group of answers.you can recode such frequent group and generate new varible in SPSS and it can be used for coding type of family as you said in the mail. I think,this will helpful for you to do further analysis of your data..you can also wait for reply from some others…good luck!..
this is what i m doing now, thanks for your comment and reminder. As we concerned on family stereotyped problem,we didn’t wish the family identify themselves as single family or what else. given that it’s cultural problem. the research team also know it could be easier to make the analysis. hence, this is the way we finally pick and let respondent to tell the household member instead. of course, when we find the coding the problem, we just wish to have a quicker way alternatively we have to do that manually.