Revisions for physical abuse estimates

30 April 2020

Estimates of physical abuse by someone under the influence of alcohol or illicit drugs have been revised following a review of the methodology for their calculation in 2019. The revised estimates have resulted in slightly fewer people reporting that someone under the influence of alcohol or illicit drugs had physically abused them.  

Logic edits are applied to the NDSHS data to resolve conflicts in responses to questions arising from inconsistencies in how people respond to questionnaire instructions and skip patterns. These are particularly an issue for questions completed in paper mode as skips are built-in and automatically applied in the online and telephone modes.

In the NDSHS, respondents are asked if they have been physically abused in the previous 12 months by someone under the influence of alcohol or illicit drugs. They are subsequently asked if any of the incidents of physical abuse involved sexual abuse.

A conflict in responses results when a respondent answers ‘no’ or does not answer the question about physical abuse in the previous 12 months, and answers ‘no’ to whether that incident of physical abuse involved sexual abuse. If they answer this secondary question as ‘no’, there is some uncertainty around whether an incident of physical abuse had occurred. Previously, to resolve this conflict, the logic edit recoded their response to the first question to record ‘yes’ for physical abuse. 

In 2019, following a review of the methodology and consultation with the Technical Advisory Group, a decision was made to revise this edit and the response provided at the sexual abuse question was changed from ‘no’ to ‘not relevant (not physically abused)’ where the respondent answered ‘no’ to the first question about physical abuse or to ‘missing/ not answered’ if no response was provided to the physical abuse question. The response to the first question was only recoded to record ‘yes’ for physical abuse if the response to the second question was ‘yes.’

These revisions ensure that the application of the edit methodology is consistent across all modes of completion of the survey and gives primacy to responses to the first question rather than the secondary dependent question.

The change in methodology for calculating these estimates is applicable to data from the 2010, 2013, and 2016 surveys. Revised estimates relating to physical abuse will not match previously published data for 2013 and 2016.  Revised estimates are not available for 2010 and the previously published 2010 estimates should not be used.

For further information about previous updates, corrections and superseded versions - refer Notes.