New ways of measuring energy poverty: Moving beyond temperature sensor data to assess and measure cold housing

Many people live in cold homes that are hazardous to health. In the absence of high-quality observed data, researchers have measured cold housing and assessed its prevalence using secondary proxy indicators. Proxy measures previously used in literature include self-assessed warmth of the home, perceived energy affordability, financial inability to heat the home in winter, and local climate zone. Using matched in-home temperature sensor data from 502 Australian homes, we assess the validity of these proxy measures by estimating the degree of association with measured indoor temperature. We also examine twelve correlated socio-demographic characteristics to explore promising alternative proxy measures.

Self-assessed perception of home warmth was shown to be the best existing proxy indicator of cold indoor air temperature (OR 2.5, CI 1.4 to 4.3), with climate zone (OR 2.4, CI 1.6 to 3.8) also shown to be a strong measure. Perceived energy affordability (OR 1.1, CI 0.7 to 1.9) and financial inability to adequately heat the home (OR 1.0, CI 0.6 to 1.6), were shown to be unsuitable proxy measures. Of the correlated socio-demographic characteristics, heating appliance type (electric heater OR 3.0, CI 1.4 to 6.2), household structure (living alone OR 2.5 CI 1.2 to 5.5), built date (built <1990 OR 2.11, CI 1.38 to 3.23) and flooring type (timber floor OR 1.99, CI 1.23 to 3.22) were strong indicators of cold indoor temperatures, and would make sound proxy measures. Our assessment of the reliability of existing and potential proxy measures of cold home temperature suggests a need to carefully select proxies, based on their known or established validity.