I think these are valuable comments, and you are absolutely correct. Limited time meant that I (1) was very short-hand in how I aggregated effect sizes/results from academic studies, (2) used simplistic point estimates. Ideally, I would have done a meta-analysis style method with risk of bias assessment etc. My main limitation is a frustrating one- time.
I did try and caveat that with trying to make all my shorthands and uncertainties explicit, but I dont think I quite succeeded at that.
One area I would push back on is the comments regarding social interventions and survey data- the methods in most/all these studies are survey effects asking women wehther they have experienced violence in the last year. To me, this seems pretty robust, and as long as the surveys are conducted to a high standard with low risk of bias (which most of the studies have dedicated sections to explain how they tried to do this, to varying degrees of success), think this is credible and internally valid data.
Hi cflexman,
I think these are valuable comments, and you are absolutely correct. Limited time meant that I (1) was very short-hand in how I aggregated effect sizes/results from academic studies, (2) used simplistic point estimates. Ideally, I would have done a meta-analysis style method with risk of bias assessment etc. My main limitation is a frustrating one- time.
I did try and caveat that with trying to make all my shorthands and uncertainties explicit, but I dont think I quite succeeded at that.
One area I would push back on is the comments regarding social interventions and survey data- the methods in most/all these studies are survey effects asking women wehther they have experienced violence in the last year. To me, this seems pretty robust, and as long as the surveys are conducted to a high standard with low risk of bias (which most of the studies have dedicated sections to explain how they tried to do this, to varying degrees of success), think this is credible and internally valid data.