This study compared the measurement invariance of paper-and-pencil (PP) and web-based (WB) administration formats through a humor survey. Participants were 401 undergraduate students divided into four groups (A, B, C, and D), and each group completed one of the four testing conditions (group A: PP → PP, group B: PP → WB, group C: WB → PP, and group D: WB → WB). The WB and PP versions of the revised Multidimensional Sense of Humor Scale (Wang, Cheng, Liu, & Ho, 2011), which measure humor production, attitudes toward humor, and humor coping, were administered to the participants. The results indicated that both of the PP and WB survey formats were practically invariant. No significant differences across administration situations were found for humor production or humor coping. Interestingly, the mean score of the WB format was significantly higher than that of PP format on attitudes toward humor. The findings suggest that researchers should carefully examine the measurement invariance and the effect of characteristics of the construct on measurement results when using a WB-format instrument.