The effectiveness of a performance-based, experimental writing program was investigated in samples of fifth- and sixth-grade children. The writing program provided instruction in specific syntactic variables from Golub's formula for syntactic density. Six teachers in one school were involved in the 15-week experimental program; four teachers in another school conducted conventional language arts activities with a control group. Thirty children were randomly selected from the fifth and sixth grades to represent experimental and control populations in the analysis of pretest and posttest results from a linguistic awareness test and of writing samples gathered before and after instruction. Results indicated that the experimental group achieved significant increases in syntactic density, compared with the control group; there were no significant differences between groups in the amount of improvement on the test of linguistic awareness or in teacher ratings of the writing samples. (Author/AA)