That sounded a bit dodgy to me, so I decided to do a regression analysis to see if it was true.
Got the following data from NSW sources
1. Incidents of Domestic Violence in 2016 from BOSCAR by postcode
http://www.bocsar.nsw.gov.au/Documents/RCS-Annual/PostcodeData.zip
2. Postcode census data in 2016 via DataPacks
https://datapacks.censusdata.abs.gov.au/datapacks/
This gave me population data per postcode, median household income per postcode, Indigenous people per post code, Christian People per post code and finally Muslim people per postcode.
3. Using this info, I then created a data table which listed the domestic violence incidence per 10,000 people for each postcode. I also listed Indigenous people, Christian people and Islamic people per postcode as well.
4. I calculated the averages for each of these, including household income. Using these averages, I was able to calculate the anomalies from the average for each of these to ensure my data fits the assumptions of OLS regression
5. I then ran a linear regression with domestic violence anomaly being the dependent variable, and household income anomaly, Indigenous incidence anomaly, Christian incidence anomaly and Islam incidence anomaly.
The results are below.
Based on this data, income is actually not a statistically significant dependent variable for incidence of domestic violence, which was a surprise.
But Indigenous status, and both religions are statistically significant at the 95% confidence interval.
What is disturbing is that as Indigenous incident and Muslim incidence rise, so does domestic violence.
But for Christianity, the negative co-efficient means that domestic violence is actually less in those post codes where more people self report as Christians above the mean.
Which is certainly my experience. This might be different in other states of Australia, but if NSW is representative of Australia as a whole, I would say ABC is wrong in saying there are problems in the Christian church in regards to Domestic Violence. In fact, Christianity help reduce that incidence.
