Satisficing in attitude surveys: The impact of cognitive skills, motivation, and task difficulty on response effects

Changes in the format and wording of survey questions can dramatically alter people's responses to them, a phenomenon termed response effects. This study explored the notion that some response effects occur because people cope with the cognitive demands of survey questions by answering questions merely satisfactorily rather than optimally (a behavior termed satisficing, Krosnick, 1991). Satisficing is enhanced by high levels of task difficulty and low levels of respondent ability or motivation.;Past research supported the notion that cognitive skills, one aspect of respondent ability, may moderate some response effects. However, this research used limited measures of cognitive skills and left unanswered questions about motivation and task difficulty. A more comprehensive test of the satisficing hypotheses was conducted using better procedures to measure cognitive skills, using the need for cognition scale to measure motivation and using accuracy instructions to manipulate motivation. Also, some respondents were distracted during task performance in order to vary levels of task difficulty. The study, thus, enabled the examination of the individual and interactive influences of the three types of variables on the likelihood of five response strategies: response order effects, acquiescence, status quo, no opinion filter effects, and non-differentiation.;Cognitive skills moderated the magnitude of the five response strategies. However, different types of cognitive skills were associated with each response strategy. Low math skills were associated with stronger response order effects and more non-differentiation. Lower inductive reasoning skills were associated with more acquiescence and more status quo responses. And, lower verbal skills were associated with stronger no opinion filter effects. Contrary to expectations, higher inductive reasoning skills were associated with more status quo responses. Cognitive skills uniquely moderated the likelihood of the five response strategies over and above the influence of need for cognition, accuracy instructions, and distraction. Finally, cognitive skills interacted with need for cognition, accuracy instructions, and distraction to regulate the five response effects. Results suggested that satisficing may occur when low cognitive skills occur in combination with aspects of low motivation and/or high task difficulty. Fortunately, results also suggested that the debilitating effects of low cognitive skills may be offset by high levels of motivation and/or low levels of task difficulty. Thus, the study provided valuable insight into how sources of unreliability and invalidity in survey data may be attenuated.