pwinskill
commented
3 years ago I think I was initially just thinking of something along the lines of the SI from the 2010 Griffin paper which has:
for between round and:
for between interventions.
I am wondering whether an upset plot might be a nice way to update, something like:
Made with ggupset, showing the % of the population with none, or combinations of interventions?
Correlations are hard to understand. We should provide outputs so that researchers can easily inspect the populations that are being drawn using the correlation parameters.
What would these outputs look like? Perhaps @pwinskill has a basic idea?
NOTE: it's been recommended to keep graphs out of our core R code. So we'll probably make a vignette to demonstrate how researchers can plot.
Initial convo and example