Relationship between Language Patterns and Antisocial Personality Disorder
McMullen, Matthew (Faculty Mentor)
MetadataShow full item record
Background:Presently the diagnosis of personality disorders, particularly those that are concerned with manipulative traits such as those associated with Antisocial Personality Disorder, are extremely time consuming and variable from clinician to clinician. With the rise of social media platforms personal writings over time have become more widely available. If a new system of analysis were to be utilized as a way to help clinical assessments it could significantly reduce the time invested as well as the reliability of the data. Aim:This study was meant to investigate whether there are quantitative differences in language patterns of those identified with a personality disorder (through the use of the MCMI-III) and those without. Approach: Deidentified transcripts of the Adult Attachment Interview were formatted so that text-analysis could be run using the R-studio (Version 1.4.1103) software. The transcripts that were identified as persons with the disorder were randomly paired with those that were identified as not having the disorder. The content analysis included the complexity of the text through the use of each group's lexical diversity, lexical density, and word count. Sentiment analysis was run which assessed not only the number of positive words versus negative but also the most common words under each of those sentiments. Similarly, overarching themes can be seen when the most frequently used words of each condition are compared by themselves and then in pairs (using the bigrams data frame). Results:Based on previous research in this area, it was expected that those with the personality disorder would show themes that are more negative in nature (e.g., aggression, fear, etc.). When the sentiment analysis was run there were differences in common words based on sentiment. However, there were not significant differences when the analyses for the texts complexity were compared.