It's an assessment that is normed against the grade. So the first grade text and questions, and then the answers are compared to first grade and given in percentiles (I did write that in my post). The fifth grade text and questions, and then answers were compared to fifth grade answers. So, yes, respective grade norms to determine the percentile scores.
Thanks. I see that.
Do you know if the scores were based on reading comprehension (questions) alone, or if oral reading accuracy and rate were also included? I ask because, if the student is a fluent reader of grade 5 text, but not a speed demon (which I find to be the case with some very good readers, who read with clarity and expression, rather than blazing through without regard for the audience), he will not necessarily be any faster on grade 1 text than on grade 5 text, past a certain point. Nor will he be any more accurate, just because the text is easier (there is, after all, a ceiling on reading accuracy for fluent readers!). The same applies to pretty much all aspects of oral reading. So if, for example, the composite score is mostly based on oral reading, and only a smaller fraction is derived from the comprehension questions, I could imagine a scenario where the score changes relatively little from grade to grade, because text difficulty matters more for other children than it does for him.
I also agree that the level of engagement is another possible factor.
