To add to previous research on gender differences in mathematics ability and achievement, this study was designed to investigate gender differences in problem-solving strategies for two extended constructed-response mathematics questions in grade 3. It is a followup to a study of gender differences on constructed-response and multiple-choice items from the Delaware state assessment program. A sample of over 300 Delaware students completed tasks from the state's assessment program. A sorting guideline was used for sorting student responses, based on current research on children's thinking in the pertinent content area and also on the variety of strategies apparent in a sample of student work. The results of analyzing student response to one task, the "Eggs" task, demonstrate an interesting pattern in that more boys than girls used the most sophisticated approach, yet more boys overall were unsuccessful at accomplishing the task. The girls were more likely to use a visual, more concrete approach, and many more girls than boys did not give sufficient explanation for the strategy used to solve the problem. The results from the "Jellybean" task show that more boys (11%) than girls (7%) used the "Level B" approach, one of the sophisticated approaches, to the task. Other than that, no significant gender-specific differences in strategies were used to solve this problem. It is reasonable to say that gender differences in strategies used might have been masked by other factors such as item difficulty. The questions and their rubrics, the sorting guidelines for both tasks, and samples of student responses to both tasks are attached. (Contains 5 tables and 10 references.) (Author/SLD)