linking two variables - mothers and children

Discussion in 'SPSS' started by Mattias, Nov 3, 2011.

  1. Mattias

    Mattias Guest

    Dear Group,

    I have a question concerning how to construct a variable. We have
    conducted a survey which we administered using household, women’s and
    men’s questionnaires. We used a household roster to map household
    members but now have difficulty linking children to their mothers –
    something which is of great importance to our study. Obviously, we
    realize that we missed including a simple variable to do this and that
    we might be forced to lose cases as a result of our mistake. I am
    wondering if anyone could help us find a way to link mothers with
    their children based on the following information.

    In our roster we have, among other things, variables giving
    information on “relationship to household head”, “sex”, “age”,
    “marital status”, and lastly a variable determining whether the person
    is an eligible respondent. Eligible women (and to link mothers and
    children we are of course only interested in eligible women) are ever
    married women in the ages between 15 and 49 years and are the women
    who also were interviewed using our women’s schedule. These are the
    women we want to link to their children. Field investigators were
    instructed to record children in the roster directly succeeding their
    mothers; however, this is likely to not have been followed fully.
    Furthermore, we included a variable in our women’s questionnaire which
    determines how many children each woman has.

    Any help would be greatly appreciated!

    Best,
    Mattias
     
    Mattias, Nov 3, 2011
    #1
    1. Advertisements

  2. Mattias

    Rich Ulrich Guest

    If that is thoroughly a problem, you may have no recourse but
    to go back to the original data packets and re-enter all the data...
    assuming that that would help.
    DO you have the original data in some form that keeps a family
    together?

    You want a "household ID". Do you have some sort of sequential
    ID for all data gathered? Do you have an ID for every Field
    Investigator, a number that is constant for a household?

    - About 1970, a friend worked on a car-repairs survey for Consumer's
    Union. When the first pass at the data showed that repairs for
    Corvettes were not more expensive than average, they knew they
    had a problem. It turned out that the keypunching had failed to
    identify "car" uniquely on every card, and the repair data had been
    separated from the ID data before being re-assembled somewhat
    randomly. IIRC, the department (or subcontractor) responsible
    had to re-punch 30,000 reports.
     
    Rich Ulrich, Nov 3, 2011
    #2
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.