admin at a clinic in Motherwell, Gqeberha. mostly here for the global health threads, the TB and malaria cost numbers make more sense to me than the philosophy ones. saturdays at the football if the roster lets me.
Thanks, I.J.J., that's a better answer than I expected. The private ballot is the part I hadn't thought of. We never had anything like that at the clinic, it was always a nurse asking face to face, so whatever people were comfortable saying was all we got.
Comparing the three methods across arms seems useful in itself. If the ballot and the direct questions move apart more in the treatment group than in control, that gap is roughly the reporting effect, isn't it? Will you publish the three measures separately, or only a combined index? I'd be curious to see them side by side.
Really glad to see the threshold written down before the results come in.
One thing I'd want to understand from the design side. IPV prevalence here is self-reported, and the drama is partly about talking more openly at home. Could the programme change willingness to report as well as the violence itself? At our clinic in Gqeberha we saw disclosures go up after an awareness push, and for a while that looked like things were getting worse when it was mostly people finally saying it.
If the treatment arm reports more honestly, a real reduction could get partly hidden and you'd land under 5% for the wrong reason. Is there anything in the survey (list experiments, or a reporting-comfort question) that lets you separate the two?
Thanks, I.J.J., that's a better answer than I expected. The private ballot is the part I hadn't thought of. We never had anything like that at the clinic, it was always a nurse asking face to face, so whatever people were comfortable saying was all we got.
Comparing the three methods across arms seems useful in itself. If the ballot and the direct questions move apart more in the treatment group than in control, that gap is roughly the reporting effect, isn't it? Will you publish the three measures separately, or only a combined index? I'd be curious to see them side by side.
Really glad to see the threshold written down before the results come in.
One thing I'd want to understand from the design side. IPV prevalence here is self-reported, and the drama is partly about talking more openly at home. Could the programme change willingness to report as well as the violence itself? At our clinic in Gqeberha we saw disclosures go up after an awareness push, and for a while that looked like things were getting worse when it was mostly people finally saying it.
If the treatment arm reports more honestly, a real reduction could get partly hidden and you'd land under 5% for the wrong reason. Is there anything in the survey (list experiments, or a reporting-comfort question) that lets you separate the two?