Using Bayesian Profile Regression to identify groups of lifestyles associated with different types of cyber abuse victimization

STATISTICAL SOCIETY OF AUSTRALIA QUEENSLAND ORDINARY BRANCH MEETING

TIME: 5:00 pm for Pizza and drinks followed by talk at 5:30, Tuesday 22 May 2018

COST: Free

VENUE: Queensland University of Technology, Gardens Point, Z Block, The Gibson Room 1064

 TITLE:  Using Bayesian Profile Regression to identify groups of lifestyles associated with different types of cyber abuse victimization

SPEAKER:  Zarina Vakhitova, Griffith University

ABSTRACT:

Recent research suggests that cyber abuse is a complex phenomenon, which could involve direct, indirect and mixed methods of abuse.  This study examines whether victims’ lifestyles and routine activities can help explain the risk of victimization from different types of cyber abuse and non-victimization. Online survey with a nationwide crowdsourced sample (N= 1,463) collected through Mechanical Turk and Bayesian Profile Regression analysis were used to answer the research question. The recently developed R package PReMiuM was used to fit the Bayesian Profile Regression. Unlike conventional regression analysis, it also allows estimating the simultaneous effect of many highly correlated variables, while handling missing data. Bayesian Profile Regression analysis identified five population subgroups based on their lifestyles and routine activities in terms of the associated risk of personal victimization from different types of cyber abuse.  The paper discusses the differences in lifestyles and routine activities between the identified groups. The advantages of using Bayesian Profile Regression over other clustering methods will also be discussed.

BIOGRAPHY:

Zarina Vakhitova is a PhD student (Griffith University) specialising in criminology and criminal justice. She is also a researcher working for Queensland Police Service. Her research interests include environmental criminology, situational crime prevention and cybercrime, quantitative methods of social research and Bayesian statistical analysis. She is published in Crime Prevention and Community Safety, Journal of Contemporary Criminal Justice, International Review of Criminology and others.

 

Membership

Keep your community strong, SSA provides a place for networking, opportunities for professional development and accreditation. In a world of alternative facts, professional bodies play an increasing important role in sharing and dissemination of trusted information.  You can make a difference, get involved in your local branch, Student membership is only $20. For more detail click here. http://www.statsoc.org.au/membership/types-of-membership/

 

 

 

Follow

Get the latest posts delivered to your mailbox:

Show Buttons
Hide Buttons