[Futuregen] Examples for the possible proof of concept piece on care

[Stefan Fors]

Thank you Ricardo, I think this would be a most valuable exercise (and look forward to the discussion this afternoon). On top of the issues raised in the paper, I think that the importance of the distinctions between absolute and relative effect measures and, correspondingly, between additive and multiplicative interactions are of great importance.

One of my colleagues, Johan Rehnberg (who will most likely join the Swedish team soon), wrote a little about this in his thesis (see the attached file: page 25ff, and the last paragraph of page 44). In his example the intersection of interest is between age and income, but the same reasoning applies to other intersections.

Zoom you soon,

Stefan

-----Ursprungligt meddelande-----
Från: Futuregen [mailto:futuregen-bounces at euro.centre.org] För Ricardo Rodrigues
Skickat: den 13 augusti 2019 16:51
Till: futuregen at euro.centre.org
Ämne: [Futuregen] Examples for the possible proof of concept piece on care

Dear colleagues,

Ahead of our zoom tomorrow at 14h00 CET, we are sending you all a short example of what we had in mind regarding a possible proof of concept exercise for care on the methods to capture intersectionality between income and gender.

The example is actually best exemplified by the paper we attach (unfortunately they apply different methodologies to diffeent examples):
depending on the method used, could we expect to have different results in terms of gender inequalities and what does this tell us about the best method to analyze these in care (one of the aims of WP3/FUTUREGEN)?

Effect modification (i.e. running regressions separately for men and
women) is akin to testing if effects of income are different for men and women. Interactions basically tell us whether the gender effect changes across a 2nd variable (e.g. income), and gives us the combine effect of gender and income. Mediation provides the direct of gender on care and indirect effect of gender on care through income (e.g. this approach for example allows us to test whether there is a gender effect at all after including the mediator).

The purpose is to inform researchers and policy makers on the dangers of interpreting results that do not accurately reflect the real dynamics between predictors and proposing a gold-standard for analyses.

Was our thinking a bit clearer after this...? We can discuss it further tomorrow.

Best wishes,

Ricardo



När du skickar e-post till Karolinska Institutet (KI) innebär detta att KI kommer att behandla dina personuppgifter. Här finns information om hur KI behandlar personuppgifter<https://ki.se/medarbetare/integritetsskyddspolicy>.


Sending email to Karolinska Institutet (KI) will result in KI processing your personal data. You can read more about KI’s processing of personal data here<https://ki.se/en/staff/data-protection-policy>.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Rehnberg_thesis.pdf
Type: application/pdf
Size: 1755542 bytes
Desc: Rehnberg_thesis.pdf
URL: <http://lists.euro.centre.org/pipermail/futuregen/attachments/20190814/8dc2161f/attachment-0001.pdf>