To test the influence of the defence norm on low-status respondent when questioned
by middleclass interviewers, a cross-section of Detroiters were asked their views
concerning two mutually contradictory propositions used at widely separated points
in the interview. As predicted, the norm led in nearly 8 per cent to agree with both
statements despite their highly contradictory character. This raises serious questions
concerning the validity of the A-scale and concerning the interpretation of the F-scale.
This study serves as yet another reminder that the research interview invariably
creates a social relationship with consequences of importance forthe interpretation