Buildings generate nearly 30% of global carbon emissions, primarily due to the need to heat or cool them to meet acceptable indoor temperatures. In the last 20 years, the empirically derived adaptive model of thermal comfort has emerged as a powerful alternative to fixed set-point driven design. However, current adaptive standards offer a simple linear relationship between the outdoor temperature and the indoor comfort temperature, assumed to sufficiently explain the effect of all other variables, e.g. relative humidity (RH) and air velocity. The lack of a signal for RH is particularly surprising given its well-known impact on comfort. Attempts in the literature to either explain the lack of such a signal or demonstrate its existence, remain scattered, unsubstantiated and localised. In this paper we demonstrate, for the first time, that a humidity signal exists in adaptive thermal comfort using global data to form two separate lines of evidence: a meta-analysis of summary data from 63 field studies and detailed field data from 39 naturally ventilated buildings over 8 climate types. We implicate method selection in previous work as the likely cause of failure to detect this signal, by demonstrating that our chosen method has a 56% lower error rate. We derive a new designer-friendly RH-inclusive adaptive model that significantly extends the range of acceptable indoor conditions 2 for designing low-energy naturally-conditioned buildings all over the world. This is 1 demonstrated through parametric simulations in 13 global locations, which reveal that the 2 current model overestimates overheating by 30% compared to the new one.