Stress is a highly subjective condition and may largely vary in different contexts. Bio-signals have been widely used by researchers and practitioners to monitor stress levels. Consequently, various bio-signals datasets for stress recognition have been recorded. The most of publicly available physiological datasets have been emotionally annotated in a context where users have been exposed to intense stressors, such as movie clips, songs, major hardware/software failures, image datasets, and gaming. However, it remains unexplored how effectively such datasets can be used in different contexts. This paper investigates the performance of the publicly available dataset named WESAD (Wearable Stress and Affect Detection) in the context of UX evaluation. More specifically, skin conductance signal from WESAD was used to train four machine learning classifiers. Regarding the binary classification problem (stress vs. no stress), models’ accuracy was rather high (at least 91.1%). However, it was found that their effectiveness in assessing stress in the context of UX was rather poor when a new bio-signals dataset was used.