Fully agree. In exactly the same way that making all of the material (protocol, data, code) available for reproducing research also is no guarantee of "better" research. Transparency is (largely) good, but can't delude ourselves that improved transparency equals improved quality.

Obviously not sufficient for causal interpretation. Just wanted to highlight some good journalism here.

Is there a live stream or recording?

I reviewed this applied example applying `synthpop` to longitudinal data. Could be of interest for you. wellcomeopenresearch.org/articles/9-57

Great post and needed (especially in epidemiology / public health where there is a lot of 'only an association BUT we should do X'). Agree with ☝️ about 'doing CI' and would prefer clarity in aims ('to estimate the effect of A' plus a serious discussion of credibility of assumptions needed for CI.