ABSTRACT
Survey questionnaires can be powerful instruments to directly measure accounting-related phenomena. Yet, due to the potential for measurement error, many economics-based researchers do not believe that results based on survey evidence are valid (Bertrand and Mullainathan 2001). These researchers argue that perceptual responses to survey questions contain significant measurement error and thus, may not capture the underlying construct of interest. This chapter makes some recommendations on the design of survey instruments to allow behavioural researchers in accounting to gather survey evidence while reducing measurement error, defined as the difference between a measured value and its true value.
