Given the sensitive nature of ethics research, the presence of social desirability bias (SDB) threatens the validity of research findings. As ethics studies often overlook this bias, we aimed to provide evidence that SDB varies across individual and situational factors. We thus investigated the influence of socio-demographic factors and survey modes on SDB. A total of 348 working adults were randomly chosen to participate in either an on-line or off-line survey containing eight versions of the Marlowe-Crowne Social Desirability (MCSD) scale. The reliabilities for the eight versions ranged from 0.35 to 0.81. Statistical tests revealed that different socio-demographic factors influence different versions of the MCSD scale. The results also showed that using on-line surveys minimizes SDB. This study provides practical implications and suggestions for future research.