Assessment results were examined for 2,351 students in a large Southwestern school district over a 4-year period. Assessment in the first year consisted of the full battery of the Iowa Tests of Basic Skills (ITBS) administered to all third graders in the district. In the following year (grade four), the same students participated in a state-mandated Writing Portfolio Assessment (WPA) program. In the third year, the fifth-grade year, the ITBS was administered and in the fourth year (grade six) the writing portfolio was administered using new prompts. Patterns of relationships were studied both within and between assessment methods. Results showed moderate to high predictive validities for the ITBS and low predictive validities for the WPA. Application of a longitudinal structural equation model indicated that later achievement (either ITBS or WPA) was related to prior achievement as measured by the grade-three ITBS, but not as measured by grade four WPA. In fact, relatively little variance in student WPA scores was accounted for using information from other assessment measures or occasions. An appendix presents the grade four and grade six writing prompts. (Contains 1 figure, 4 tables, and 17 references.) (Author/SLD)