Determination of the intentions of the test developer is fundamental to the choice of the analytical model for a rating scale. For confirmatory analysis, the developer's intentions inform the choice of the general form of the model, representing the manner in which the respondent interacts with the scale; these intentions also inform the choice of the precise statement of that form, representing the intention of the analyst to construct, for example, an "equal-interval" scale. The nature of the Likert rating scale is discussed. Examples of general forms and precise statements are given. Forms of the measurement model discussed include the dichotomous case, the Andrich model for holistic scales, the Glas model for incremental scales, and the McCullagh model for incidental scales. Means of modeling and communicating the intentional form of the scale are outlined. Other issues addressed include incorporating the developer's intentions, anchoring the scale calibrations, and general structural concepts. A sample design problem is presented. Two tables and 16 figures are included. (TJH)