Senior managers’ mental models of safety culture: An application of linguistic analysis

This research investigates the suitability of linguistic analysis to identify senior managers’ mental models of safety culture. Schein (2004) defines culture as consisting of artefacts, beliefs, values and underlying assumptions, with language as one of these artefacts. Popping (2000) describes the analysis of language as suitable to investigate mental models. Although frequently used in the culture research (e.g. Giberson et al., 2009), questionnaires have been criticised as being more suitable to assess organisational climate, not culture (Mearns, Flin, Fleming & Gordon, 1997). Senior managers have been identified as a main driver of organisational safety, but how they perceive safety is still not completely understood (Flin, Mearns, O’Connor & Bryden, 2000). Accordingly, the present research investigates language to gain insights of senior managers’ mental models of their organisation’s safety culture. This is done in air traffic management (ATM) which can be considered a high reliablity industry (Weick, 1987).
A sample of N=9 senior managers from three ATM organisations participated in semi-structured interviews. Participants’ responses to the question “How would you describe the safety culture of your organisation?” were analysed using the Leximancer software (Smith & Humphreys, 2006). Leximancer automatically identifies concepts in a text corpus and analyses how these relate to each other using Bayesian logic (Smith, 2005). Resulting concepts are displayed in form of an insights dashboard reflecting their frequency of co-occurrence with safety and the strength of their relation to safety (meaning how frequently they occur without safety. The dashboard can be viewed as reflecting senior managers’ mental models of safety culture. Included concepts suggest that senior managers, when talking about safety culture, most frequently mention ‘culture’, ‘reporting’ and ‘seriously’ together with safety. It further suggests that they least frequently mention ‘units’, ‘understand’ and ‘controllers’ in relation to safety, when describing safety culture.
Although still at an explorative stage, the present findings support the suitability of this form of linguistic analysis to understand how language reflects organisational culture. Through its visual display of the safety culture components this form of analysis could support organisations to reflect about strengths and weaknesses of their culture.