Hi all,
First post so any comment/suggestion is welcome.
I'm using 7 consecutive censuses (1851-1911) for a specific region. Links have been made with registers, therefore I know almost every kin relationship in each census (97% individuals linked and 99% households linked in 1891 only). My research interest is to assess the existence of residential patterns at the family level differentiated by socioeconomic context (fringe, rural, saturated rural, urban). At the end of the period, the region is rapidly modernizing and I want to test the presence of an urban effect on residential choices through time.
I was aiming to use a dyadic dataset stratified by relationship (parent, children, sibling, grandparent, grandchildren, uncle/aunt, nephew/niece, and cousin). I have a hard time finding a model that would fit my requirements. Note that people can be part of multiple dyads for a single census year. Also, it is not uncommon to find people with a hundred cousins in a census...I was wondering if I should weigh the relationships.
I would like to test for lineage (paternal or maternal) as well as age (and/or age difference), being head of a household or not, death of a parent, and location of course.
I read about APIM models, but I think the melogit model could get me there. I was looking for any advice on the design of the modelization of my data. I think it is pretty interesting, I'm just not sure how to manage all that.
Thanks for your time and consideration.
JS
First post so any comment/suggestion is welcome.
I'm using 7 consecutive censuses (1851-1911) for a specific region. Links have been made with registers, therefore I know almost every kin relationship in each census (97% individuals linked and 99% households linked in 1891 only). My research interest is to assess the existence of residential patterns at the family level differentiated by socioeconomic context (fringe, rural, saturated rural, urban). At the end of the period, the region is rapidly modernizing and I want to test the presence of an urban effect on residential choices through time.
I was aiming to use a dyadic dataset stratified by relationship (parent, children, sibling, grandparent, grandchildren, uncle/aunt, nephew/niece, and cousin). I have a hard time finding a model that would fit my requirements. Note that people can be part of multiple dyads for a single census year. Also, it is not uncommon to find people with a hundred cousins in a census...I was wondering if I should weigh the relationships.
I would like to test for lineage (paternal or maternal) as well as age (and/or age difference), being head of a household or not, death of a parent, and location of course.
I read about APIM models, but I think the melogit model could get me there. I was looking for any advice on the design of the modelization of my data. I think it is pretty interesting, I'm just not sure how to manage all that.
Thanks for your time and consideration.
JS
