This paper examines aspects related to the validity of process-product research involving classroom observation and teacher effectiveness. Differences in categories, terminology, and definitions of five classroom observation instruments are delineated and discussed in relation to their various effects on the validity of process-product research findings. The instruments analyzed are: (1) Teacher and Child Dyadic Interaction observation system; (2) Reading and Mathematics Observation System (RAMOS); (3) Coding System for the First Grade Reading Group Study; (4) Classroom Observation Instrument; and (5) Group Reading Interaction Pattern Observation Instrument (GRIP). Inconsistencies between actual verbal behaviors and teachers' intentions for these verbal behaviors are also discussed. Suggestions for selecting and writing terms and definitions point out the necessity for consistency of terminology and definitions of categories and subcategories across observation instruments. It is recommended that coding of teachers' intended behaviors, rather than their exact linguistic form, would result in more valid findings in the area of process-product research. (Author/JD)