Research focused on linking indoor thermal conditions to human health outcomes has typically relied on temperature or popular thermal comfort models (i.e., Predicted Mean Vote, PMV). However, recent studies have found that these approaches have low predictive accuracy when applied to more recent and varied datasets and that there are likely other variables and/or characterizations that are important to health outcomes. To investigate the usefulness of other thermal metrics, we used easily-accessible measurements of temperature and relative humidity (RH) from in-room monitors in university lecture halls to calculate a suite of eleven thermal variables. We then paired this thermal data with thermal sensation votes and cognitive test scores from graduate students attending classes in these monitored spaces. Our analysis of the data revealed three key takeaways:
- By calculating the suite of eleven variables, we’re actually measuring three things: variation in indoor temperature, variation in indoor moisture, and variation in indoor-outdoor differences (see figure below).
Spearman rank correlation matrix showing eleven thermal variables clustered into three groups, further supported by factor analysis and principal component analysis. The color scale shows the strength and direction of correlations.
- Many commonly used indoor thermal characterizations can emphasize variations in temperature and not moisture. For example, the PMV estimates in this study show strong positive correlations with temperature and low correlations with RH; calculated heat index tracked with indoor temperature; and the use of temperature and RH as separate independent variables in mixed effect regression models emphasized an association with temperature and not moisture.
- Calculated indoor enthalpy could be a useful metric for better understanding the combined effects of temperature and moisture on human health outcomes. Enthalpy correlated with indoor temperature and moisture in the most balanced way and showed evidence of distinct associations with cognitive test scores and thermal sensation votes.
Read the full paper here: https://doi.org/10.1016/j.indenv.2025.100098