This paper presents a model for systematic instructional design that includes mental model analysis together with the procedures used in developing computer-based instructional materials in the area of statistical hypothesis testing. The instructional design model is based on the premise that the objective for learning is to achieve expert-like mental models, and instruction should be designed to help learners build relevant mental models in the specific domain. (The term mental model is defined as a person's internal, domain-specific representation that may be incomplete or unstable, and the term relevant mental model is defined as an internal, domain-specific representation that is relevant and useful for a person's subsequent understanding of and problem solving in the field.) It is proposed that mental model analysis be integrated into the design of instructional materials so that cognitive task analysis can be used to produce effective instructional strategies. The three phases in the design of instruction covered by the model are described: (1) analysis of instructional outcomes; (2) development of instructional material; and (3) implementation, evaluation, and revision of the instructional material. Seven procedures for applying this model are discussed in more detail: (1) identify instructional goals/objectives; (2) conduct mental model analysis; (3) identify the learners' entry level knowledge/skills; (4) develop instructional strategies considering mental models; (5) develop instructional materials and tests; (6) formative evaluation; and (7) revision of the instructional program. A discussion of the instructional effects which resulted from the application of the mental model strategies in an introductory statistics course concludes the paper. (2 figures, 32 references) (BBM)