Hi everyone
I am researching about depressive symptoms and gender. Specifically, the asociation between marital status, head of household, number people in the home and ocupation (employed, unemployed and housework). I did make univariated and multivariate analysis but someone suggested me that i make typologies / composite measure, for example:1) single + alone + unemployed, vs 2)divorced + >3 people and house work, i could get better results.
do you know some paper about this topic? what have been your experiences with this method?
thanks for your help
Salutes
In case you haven´t checked it out yet.
Greetings =)
Scand J Public Health. 2015
Marital status, labour force activity and mortality: a study in the USA and six European countries.
https://www.ncbi.nlm.nih.gov/pubmed/25868643
In case the data is sparse, multivariate bayesian techniques work really well. This has been my experience. If you have a large sample then typologies and multivariate techniques may converge. Since, you are dealing with a social behavior some mathematical models from psychology like prospect theory may also be of help if you know how each variable contributes to the final inference.
Dear Gabriel,
It would be helpful to know how you are measuring depressive symptoms. For example, if this is a single scale variable then it could be treated as a dependent variable in a multiple linear regression model. Or if it is a binary variable then it could be the DV in a binary logistic regression model. Each of these models would retain the complexity of the individual variables, suggesting which ones are significant.
I would personally recommend this before considering typologies as otherwise the choice of typology would seem to be arbitrary.
@Shameek Bhattacharjee thanks for your answer. The size of my sample are 930 men and 1200 women. i dont know prospect theory i will review it.
@Peter Samuels the measure a a standarized clinical interview, it gives a dicothomic outcome. Thanks for your advice, i thinks typologies are arbitrary too.
- Badges
- Science topic