Academic journal article Advances in Mental Health

Lexical Analysis of Coronial Suicide Reports: A Useful Foundation for Theory Building

Academic journal article Advances in Mental Health

Lexical Analysis of Coronial Suicide Reports: A Useful Foundation for Theory Building

Article excerpt

As an outcome of complex circumstances and factors, suicide has been interpreted in many ways, including emphasizing the importance of social (Durkheim, 1951), medical (Dorpat & Ripley, 1960; Ernst et al., 2004), environmental (Foster, 2011) and predicament-related factors (Pridmore, 2009). Such varying perspectives have also tended to be associated with a variety of different methods, types of data and ways of analyzing those data.

In the planning of a new study on suicide, strategies for tempering the influence of some of these competing perspectives and assumptions were considered. While it was recognized that no research method, data source, or method of analysis is free of assumptions, the current study sought a data source and method that might to some extent inform this area from another perspective.

Analysis of coronial reports

In seeking appropriate data on completed suicide, the research team acknowledged the potential benefits of psychological autopsy, including richness of data (Bertolote, Fleischmann, De Leo, & Wasserman, 2004; Scocco, Marietta, Tonietto, Dello Buono, & De Leo, 2000), however, limitations of such methods were also evident including selection bias, confounding influence of extraneous variables, and lack of standardization of assessment instruments (De Leo & Evans, 2004; Pouliot & De Leo, 2006). As a result, and in recognition of precedents (Bennewith et al., 2005; Ward, Shields, & Cramer, 2011), coronial reports and findings were selected as the data source for this research.

As noted elsewhere (Shiner, Scourfield, Fincham, & Langer, 2009), coroners' reports are produced for specific purposes, under certain circumstances, and not intended for research per se. While there may be variability across reports (Bennewith et al., 2005), these sources seek to establish externally verifiable information from a legal perspective, which may be a reasonable basis for investigating common factors across numerous suicide cases (Shiner et al., 2009). Procedurally, the use of these reports for suicide research has some benefits. It is less intrusive into the lives of family members and others recently bereaved. It is cost-effective, providing considered information from a number of key informants (such as police officers, family members, experts, pathologists, etc.), who may not normally be available to researchers. Finally, coroners' investigations seek to be thorough, fulfilling numerous legal and procedural requirements, and coroner's reports reflect the systematic nature of these investigations.

Computer-assisted lexical analysis

While no method or tool is free of assumptions, and indeed subjectivity may be desirable in some instances (Patton, 2002), this project sought to explore the benefit of a more detached method as a research starting point. The researchers identified a computer application for analyzing textbased, documented material (Leximancer(TM)1) which is reported to have a substantial degree of 'objectivity' (Smith & Humphreys, 2006). Analysis through the Leximancer(TM) program is grounded entirely within the text, rather than on researcher-driven interpretive coding, so is reported to be less subjective than other qualitative analysis methods (Hewett, Watson, Gallois, Ward, & Leggett, 2009). Leximancer(TM) detects and extracts meaningful relationships among concepts using artificial intelligence. The software identifies words, phrases and concepts, and applies concept frequency and co-occurrence data using a statistical algorithm to map the main concepts in the text and their relative importance, the strengths of links between concepts (how often they co-occur), and similarities in contexts where links occur (Smith, 2005).

Leximancer(TM) has previously been used to analyze suicide-related material (Petchkovsky, Cord-Udy, & Grant, 2007), and also applied in similar larger scale studies, investigating independent reports of adverse incidents. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.