Repertoires of information and perception gap

Existing studies test the effect of media and other socio-economic factors on the perceptionof crime trends. However, many of these studies fail to account for the practical situation ofsimultaneous engagement with the media and non-media information sources. The purpose of this study is to explore repertoires of information sources and whether these and any other demographic variables relate to the perception of national crime trends for individuals who do not have any prior experience with the criminal justice system. Latent class analysis (LCA) is used to derive a small number of homogeneous groups according to their engagement with various media and non-media sources using the modal class assignment. Further, multinomial regression is used to test the effect of belonging to one of the classes on the perception of national crime trends. LCA indicates the presence of three distinct subgroups in the population corresponding to those who used only media, non-media, and both types of information sources. The study finds that majority of individuals engage with only a few selected information sources. Multinomial regression reveals that the choice of information sources and the perception of national crime trends are not independent. Those with larger repertoires are likely to perceive a high rise in the crime trends nationally and are more pessimistic than those who use selected sources. Engagement with multiple information sources contains behavioural information about our perception. A larger repertoire of information is likely to have a detrimental effect on our perception of national crime trends