The intention of this article is to examin the nature of cognitive representations that respondents create and utilize when making evaluative judgments in surveys. The following two hypothetical statements are the starting point of the analysis: (1) It is theoreticaly appropriate to view the latent variable as being unidimensional continuum; (2) The item order affects both, the cognitive representation of the underlying dimension and the factor structure of the answers. Experimental results confirm both statements although the content of measured concepts and the context of the survey should be considered.
|
