New working paper from our Predlife consortium #sociology#causalinference#bayes#stats#rstats
priorsense package for prior and likelihood sensitivity analysis is now on CRAN. It works with Stan (brms, rstan or cmdstanr) fits, or just posterior draws. See more in Noa Kallioinen's post at discourse.mc-stan.org/t/priorsense...#Bayes#MCMC#rstats
priorsense 1.0 has been released and is now available on CRAN. priorsense is an R package for efficiently checking whether the posterior is sensitivity to changes to the prior or likelihood (see the ...
I will discuss Bayesian causal inference for panel data using R package dynamite in StanCon2024, held on 10-12 September at Oxford. #causalinference#bayes#statsmc-stan.org/events/stanc...docs.ropensci.org/dynamite/
Another though: Depending on your situation, you could perhaps define this as a multivariate response consisting of three binomial variables so that you first estimate number of success in the first level, and the number of failures from that is the number of trials for second level and so on? 3/3
But if you are working on software which allows constraints/priors on model parameters, you could use same predictors for each level but constraint some coefficients to zero. 2/
Technically I think it would be possible, and we actually entertained adding this possibility in github.com/ropensci/dyn..., but eventually decided not to implement this as the interpretation of the model and defining the reference level would be tricky in general. 1/
New paper by my PhD student Lauri Valkonen et al. on combining experimental and observational data for causal inference in pricing decisions: muse.jhu.edu/pub/56/artic...#stats#causalinference#bayes
š Join us at the Finnish Health Economics Seminar Series on Monday 3 pm for a presentation by Thang Dang from Norwegian Institute of Public Health with a title " Language Training, Refugeesā Healthcare Integration, and the Next Generationās Health". #HealthEconomics
New paper "Estimating causal effects from panel data with dynamic multivariate panel models" published in Advances in Life Course Research: www.sciencedirect.com/science/arti...#causalinference#bayes#stats#lifecourse#sociology@invest-flagship.bsky.social
Panel data are ubiquitous in scientific fields such as social sciences. Various modeling approaches have been presented for observational causal inferā¦
Bayesian statsky friends! A reviewer is complaining that we used a Bayesian model on a full population of data (a country-week panel from 2020āmid-2021), and that Bayesian methods can't be used there. That's wrong, but is there a citation we can use to say it's fine to do this with IR panel data?