Interviewers/askers skew answers because their personal characteristics (e.g., ethnicity, gender), manner of speaking, types of questions asked and not asked, and so on, affect interviewees’ responses. Here’s Jonah Goldberg’s comment about how interviewer Brianna Keilar affected Hillary Clinton’s answers about her emails: “After ducking the press for months, Clinton sat down for an interview with CNN’s Brianna Keilar. . . . Keilar covers the Clinton campaign and has every incentive not to offend her famously vindictive sources.” So, Kevlar retained access to Hilary, Bill, and members of “Team Clinton” by asking Hilary mild, even “muddled” queries and “avoid[ed] asking meaningful follow-ups”; thus, Hilary could make a “plausible defense” for her “stealth server” and emails. (Chicago Tribune, 7/13/15, sec. 1, p. 15)
For a complete discussion of the myriad ways askers skew answers–including probes, bribes/incentives, and props–see Part Five, “Askers”, in my book, The Problem with Survey Research, pp. 197-278.