Bivariate probit model

Hi to everyone,
I am working on climate change and education. I want to assess the effects of climate change on education through health channel. For this, I want to a bivariate probit model. Here is the equations:
education= f(health, household and child caracteristics) (1)
health= f(climate change, mother’s caracteristics).
I would like to have to idea about it.

I think your question is about variables to measure climate change, education, and health. Please let me know if that is not the case. I am not sure which IPUMS datasets you would like to use or what your focal population is; however, all IPUMS microdata sites include topical drop-down menus to help researchers locate variables of interest based on broad topics. To get data on climate change, education, and health, my best recommendation is IPUMS DHS. IPUMS DHS includes contextual variables that should help you capture some aspects of climate as well as measures of education and many variables related to health. IPUMS Terra integrates population and environmental data, though I am not aware of health variables included in these data. IPUMS PMA and IPUMS International both include variables related to water source, but I don’t know that they would be adequate for your research application; IPUMS PMA will have much more robust coverage of health measures. I am not aware of measures in other IPUMS datasets that are well suited to explore climate change.

If you can provide more detail about the IPUMS dataset you are using, I may be able to make more targeted suggestions.

Thank you dear Williams for the reply.
I am studying the effects of climate change on education through the health channel. For this, I am using data on education from IPUMS international. I extract data on school attendance, educational attainment and disability, which is the proxy of health status.

What do you think about merging education data from IPUMS International and health data from IPUMS-DHS? Only that the IPUMS-DHS database contains children under 5 years old. Could using the DHS fever variable to study the effects of climate change on the education of children aged 6 to 18 can be a problem?
Thank you for support.

There is no way to link the individuals in the IPUMS International data with those in the IPUMS DHS data. However, researchers can link the two datasets at the first or second administrative geographic unit to include summary information from the census data in their analyses of DHS data. Presumably you could do the inverse and create summary measures from the DHS data to include in your analysis of IPUMS International data as well.

Thank you very much.

1 Like