Cold and warm stimuli delivered over a larger skin area tend to be more easily detected/elicit stronger sensations, a phenomenon referred to as spatial summation. The aim of the present study was to clarify how stimulation area affects thermal detection processes by evaluating whether increasing the stimulation area simply reduces the detection threshold or also reduces the uncertainty of the detection process. Psychometric functions were fitted to the detection performance of 16 healthy subjects. Stimuli (duration: 200 ms; rate of change: 300 °C/s) were delivered to the volar forearm using a Peltier-effect contact thermode and three different stimulation surfaces (23 mm², 69 mm², and 116 mm). Stimulation intensities were selected trial-by-trial by the psi marginal method to optimize estimation of slope and threshold parameters of the psychometric function. The raw data (100 stimulus–response pairs per subject per surface and per modality) was used to fit group-level hierarchical models of cold and warm detection, allowing to assess the effect of stimulation surface and account for inter-individual variability. Increasing stimulation area led to a compression of the psychometric function towards baseline skin temperature (reduced threshold and steeper slope), suggesting that spatial summation reflects a change in the precision of the neural representation of the stimulus which in turn influences the ability of the nervous system to distinguish true stimuli from sensory noise. Regardless of area, with the stimulation settings used in this study, cold detection appeared easier than warm detection, possibly because of structural and functional differences between cold- and warm-sensitive afferents.