First, your good idea to compare your son's scores year-to-year is about to collide with a statistical hazard. The tests at each grade level are not truly comparable. So make comparisons with caution.

Second, the scores of many standardized test have a large "margin of error." That means if the same child takes the test three times in a month, his scores could easily vary by as much as 20-40 points (measured by the national percentile rank).

Here's a suggestion. Instead of looking at your child's national percentile rank (NPR), I suggest you look at the raw score — indicated as "R.S." on your printout. Then ask the school how many items were on each subtest as the number varies from grade level to grade level, test to test. Use that value to compute your child's percentage of correct answers for each subtest. You may be surprised to see that your child is getting an equal or greater number or questions correct even though his NPR varies.

If you want a better measure of progress, year to year, I recommend using reading assessments such as an IRI (Informal Reading Inventory), especially for younger students. Most schools assess students using this type of instrument, as well.

Remember, the SAT-9 is only one indicator of performance, and a highly imperfect measure at that. If you want a truer picture of your child's academic progress, use the SAT-9 only in combination with other measures.