Different UX questionnaires utilize various types of items, e.g., rating scales or semantic differentials. Another difference in UX questionnaires is item polarity. Some questionnaires maintain a consistent polarity with positive terms on the right, while others randomly change this position. Studies have shown that item polarity does not significantly impact the results if statements are used as item format. However, such investigations are currently lacking for semantic differentials. We investigate in two studies whether changing to a one-sided polarity of semantic differential items influences the effort required to answer the items and the measured scale scores. In the first study, we used the User Experience Questionnaire (UEQ) to evaluate Microsoft PowerPoint, WhatsApp and Google Maps with the original UEQ and a modified version with one-sided polarity. In the second study, we investigated again WhatsApp with the HED/UT and Trustdiff questionnaires. Results showed that the effort to complete the questionnaire was lower for the one-sided polarity version, but differences were small and practically not relevant. The impact of item polarity on scale scores was low, only for one of the UEQ scales there is a significant difference. Scale consistency was slightly higher with one-sided version, likely due to a small answer tendency introduced by the one-sided polarity. Overall, the impact of item polarity on the results of semantic differential scales is low, but differences can arise depending on the scale’s semantics and the investigated product. Comparisons of questionnaire versions that do not have the same item polarity should be done carefully.