A concept-map assessment consists of a task that elicits structured knowledge, a response format, and a scoring system. Variation in tasks, response formats, and scoring systems produce different mapping techniques that may elicit different knowledge representations, posing construct-interpretations challenges. This study examined two mapping techniques: (1) students generated the 10 concepts from chemistry to construct a map; and (2) the assessor provided a list of 10 concepts. Two concept-lists were randomly sampled from the domain to examine the effect of concept sampling on map scores. Forty high school students, two teachers, and one expert participated. Results indicate that: (1) the two mapping techniques were statistically equivalent; (2) students' concept-map scores generalized across samples of concepts; (3) concept maps could be reliably scored, even though they involved complex judgments; and (4) multiple-choice test and concept maps measure somewhat different aspects of science knowledge. Appendixes contain a discussion of compiling the list of concepts, a list of concepts considered for the three conditions, a sample of instructions, and a matrix of the relations between pairs of concepts. (Contains 9 tables and 32 references.) (Author/SLD)