Disguise and imitation of language style in WhatsApp messages

Abstract

The deliberate manipulation of evidence is routinely considered in most branches of forensic science, but in the field of forensic authorship analysis it is a relatively new concern. This paper reports the findings from an experiment which explored what happens when authors are asked to manipulate their language style. We asked 120 Dutch students to write WhatsApp messages in their normal style and in two manipulated conditions: disguise and imitation. Results show that authors are able to introduce major changes in their language use, which are most substantial when they have access to messages written by the person they are trying to impersonate. These findings imply that the possibility of disguise and imitation should be considered when conducting forensic authorship analysis, especially when more than one candidate for authorship could have had access to the same device or account.