Topic Modeling as a Tool to Analyze Child Abuse from the Corpus of English Newspapers in Pakistan

Abstract

Child maltreatment is a global issue deeply rooted in cultural, economic, and social practices. The present study investigates the representation of child abuse in the mainstream newspapers of Pakistan (i.e., Dawn and The News) from 2013 to 2020. This specific time frame is that these cases increased during these years. The data regarding the child abuse cases were accessed from the LexisNexis database. A computational technique of topic modeling (Latent Dirichlet Allocations) was applied to discover the hidden topics from the filtered data. The study underpinned Computational Grounded Theory as a theoretical frame for this study. The results reveal topics related to awareness of child abuse, the types of child abuse, and the law and procedural matters specific to child abuse cases in Pakistan. The study brings possible implications that child protection laws are to be improved to overcome the menace of child abuse in Pakistan.