B4690 - Accounting for selection bias a study of Domestic Violence Abuse DVA in the Avon Longitudinal Study of Parents Children - 15/09/2024

B number: 
B4690
Principal applicant name: 
Annie Herbert | University of Bristol (United Kingdom)
Co-applicants: 
Rachel O'Donnell, Professor Laura Howe, Professor Christine Barter
Title of project: 
Accounting for selection bias: a study of Domestic Violence & Abuse (DVA) in the Avon Longitudinal Study of Parents & Children
Proposal summary: 

Previous research has suggested that young people who grew up in homes where their parents were violent or abusive towards each other are more likely to have violent or abusive relationships themselves. We have been looking into this in Children of the 90’s data, to get a better understanding of the relationship between growing up around violence or abuse between parents in the UK, as reported by the parents at the time of the violence or abuse, and being in a violent or abusive relationship as a young adult. However, some participants (both parents and children) will have left the study since being initially recruited, and it is likely that those who are in abusive relationships are more likely to be the ones that leave. In this study, we want to find ways to account for these people that have left, so that our estimates about the risks of being in an abusive relationship are more accurate.

Impact of research: 
These findings can inform future longitudinal work in DVA, both in parents and young adults, in terms of a better understanding of the sensitivity of findings to loss-to-follow-up, and publicising methods to correct for this loss-to-follow-up, within the DVA research literature. We expect the findings will be submitted to a relevant journal, such as Trauma, Violence & Abuse. By the end of the project, the student will have knowledge and skills in ALSPAC data, selection bias, regression analyses, multiple imputation, and inverse probability treatment weighting using the Stata or R programming package (depending on preference), as well as experience in interpreting study findings for public health.
Date proposal received: 
Friday, 13 September, 2024
Date proposal approved: 
Sunday, 15 September, 2024
Keywords: 
Epidemiology, DVA, Statistical methods, Cohort studies - attrition, bias, participant engagement, ethics