The cognitive response model for survey response, usually attributed to Tourangeau (1984), generated anentire field of research to study and reduce measurement error by evaluating and improving survey questions. While the bulk of this research has been undertaken and reported relative to household and general population surveys, this model has also taken hold in establishment surveys. However, characteristics of establishments as respondents have challenged the direct applicability of the model, and has resulted in some modifications (Edwards and Cantor, 1991). This paper suggests further modifications to the basic four step cognitive model in the context of establishment surveys. Although additional steps were suggested by our research on the survey response process in large multi-unit companies (Sudman et al., 2000), we find that reinterpretation of the literature lends additional support to our proposed model. This paper will review various enhancements to the basic cognitive model, place them in the context of establishments, and provide supporting arguments for our proposed model. We will also raise several unresolved or seemingly conflicting issues that remain.