Arabella - thank you for those links (it cleared up the "prosecuted" vs "convicted" question for me in addition to being helpful in other ways. For me, Table 2 would have been useful with a "not guilty" column and some other data as the cold is making me reluctant to do even basic mental arithmetic).
I was struck by this observation in the study which, as so frequently argued on FWR, calls for the collection of accurate data:
Child abuse does not have any geographic, cultural or religious boundaries and more research is required on GLCSE across the world on national responses, and policy and practice. Jones and Florek (2015) confirm that countries may see GLCSE differently, but similar offences have triggered law and policy in the European Union, Australia and the USA. Authoritative data on the age, gender, occupation, ethnicity and religion of offenders needs to be collected for the UK from court records as a matter of urgency. The controversy related to GLCSE can be resolved through the availability of authoritative data on the identity of the offenders
As a general point (IPSO seems to have been misrepresented in their findings) I've a strong suspicion that IPSO, amongst other comparable bodies, should consider employing the services of people who are used to writing systematic reviews and meta-analyses when it comes to adjudicating matters such as this. At the very least, it would alert them to the need to be aware of checklists for assessing the quality of the studies and the relative weight of evidence and state that they've followed this in their assessments.