Thermal comfort is crucial for people well-being and productivity in buildings, and it has a strong effect on building’s energy usage. The main models for assessing thermal comfort, like predictive mean vote (PMV) and adaptive models, need many environmental measurements, making monitoring complex. That has sparked a debate: some advocate reducing the number of variables, while others stress sticking to the original models. This article evaluates current research trends and addresses the limitations of globe sensors, the most commonly used instrument for measuring mean radiant temperature. To address these is- sues, the study conducts a sensitivity analysis using latin hypercube sampling with multidimen- sional uniformity and employs data from 41 measurements conducted in educational buildings spanning different academic levels, from preschool to university. This analysis examines various simplifications in the measuring procedures for the PMV and adaptive models proposed by ASHRAE 55, ISO 7730 and EN 16798. For this purpose, two mean radiant temperature accuracies are used according to the minimum requirement and the desired value specified by ISO 7726, ±2 ¿C and ±0.2 ¿ C respectively. The results suggest that models with fewer variables tend to overestimate thermal comfort levels. However, this models are valid as their estimations fall within the uncertainty range of the complete model, as long as the mean radiant temperature accuracy is ±2 ¿C as achieved with globe sensors. When mean radiant temperature accuracy improves (i.e. ±0.2 ¿C), the outcome of the models with less variables imply an overestimation falling out of the uncertainty of the complete model.