Many have suggested that human interviewers can invoke social norms with regard to propriety in responding. We conducted a nationwide phone survey in parallel with a nationwide online survey. We selected a series of items that we expected would vary with regard to socially desirable responding (e.g. belief in God, driving over the speed limit, giving money to charity, gambling, etc.). We found that for items separately scaled as undesirable, online respondents were more likely to endorse them as being true for them. For items scaled as desirable, phone respondents were more likely to admit to them being true of them. As an example, online respondents admitted to often driving over the speed limit at a higher rate. In contrast, online respondents were less likely to admit going to church, mosque, or synagogue. We will also present results comparing the ratings of these items with the endorsements of these items in both modalities.