My research topic is " Factors effecting on infant mortality among the slum dwellers". For this I have to specify some independent variables which are significant or not for the occurrence of infant mortality ( which is dependent variable) among the sample slum dwellers. Now my question is how should I articulate independent variables in SPSS for binary? Suppose weight of the child at birth is one of the independent variables, now whether it is a significant factor or not for the occurrence of infant mortality ( in case of binary logistic regression) . So how do I put this variables in one column? Because, one household experiences not only one live birth or weight of only one child's birth , but also multiple live births say 1,2,3 etc...& among multiple births of child , may be some household experience infant mortality too. So how I just merge all type of weights of the children at births ( live birth + dead infant) in just one column?
Hello Minakshi,
The answer depends on how you are choosing cases for analysis. If you restrict your analysis to first-born child (so that each case is independent of others), then there is no special analysis set-up or adjustment needed: just enter the birthweight for the child as the IV and the survival status as the DV. The same would be true if you restricted your cases to the last-born child for each family.
The problem comes when you collect data on multiple births per family, since these are not completely independent events. You could apply a random-effects, mixed model (multilevel model) wherein births are nested within families, as well as possibly families nested within neighborhoods for your analysis.
The second approach uses more of the data, and likely gives a richer description of reality, but the first approach is much simpler to analyze. In large part, the choice depends on the specific question(s) you'd like to be able to answer.
Good luck with your work.
Sir... I'm not able to solve this problem...actually earlier I thought the term 'multiple births" in a different way..but later when I understood about this term, then things are again involved in puzzled..Sir in my analysis, I have not taken multiple births per family (like twins baby, triplets etc). Because, in my sample population, no one family experiences this cases. I have taken all live births between 2014 to 2018 per family to analysis the factors affecting on infant mortality. So without restricting my analysis only on 1st birth or last birth, is it possible to construct a column by merging all weights of the childs per family ( i.e weight of the 1st child , weight of the 2nd child etc, which are totally independent with each other that means no twins or triplets) ?
and for my analysis, is it possible to use simple binary logistic regression model in SPSS?
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS