A student has improved on some review tasks but the log still shows this issue: The student flags uncertain questions during timed work but later reviews only whether the credited letter was right, not what reasoning pattern caused the uncertainty. Which plan gives the best evidence of stable improvement?