Ten questions concerning statistical data analysis in human-centric buildings research: a focus on thermal comfort investigations

Abstract

Given the large amount of time we spend indoors, designing and operating buildings that are safe, comfortable, and conducive to productivity and well-being is essential. To achieve this goal, in the past decades, research has been conducted to investigate the influence of the indoor environment on occupants. Thermal comfort has been the subject of most investigations in this field. However, despite being a consolidated research topic since the 1920s, statistical practices for analysing thermal comfort data often rely on simplified premises, which may be due to several possible factors (e.g., limited computational capabilities and lack of training). Consequently, important aspects of data analysis are often absent or overlooked. Recent statistics and statistical software advances have provided more options for effectively modelling complex issues. However, properly using these tools requires a solid understanding of statistical analysis, increasing the risk of misuse in practice. This paper presents ten questions highlighting the most critical issues regarding statistical analysis for thermal comfort research and practice. The first four questions provide general perspectives concerning statistical data analysis, while the remaining ones address specific problems related to thermal comfort research, but that can extend to all human-centric research in the built environment. Additionally, the last five questions demonstrate the practical significance of analysis pitfalls (i.e., sampling variability, selection bias, variable selection, clustered/nested observations, and measurement error) through examples with synthetic data. This study provides insights into the current statistical ‘habits’ in thermal comfort research and, more importantly, help researchers better define and conduct their statistical analyses.