There was kind of an interesting news phenomena to watch yesterday (it may have started a little earlier). National data were released providing the ACT test score means by state. It was reported that North Dakota scores were falling (at an average of 20.7) and were below Minnesota and South Dakota. The national ACT average is 21.1. These are the statistics.
I always tell my students that to interpret research you consider the statistics in combination with the methodology.
Here is the rest of the story or here is the methodology. North Dakota pays for students to take the ACT and expects eligible students to take the exam. I saw a report from Illinois that 10 states have such a requirement (or at least there are 1o states in which greater than 90% take the exam). Illinois evidently had the top average (20.9) among this group. As a comparison, 70% of eligible students in Minnesota took the ACT.
I watched the online posted stories change across the day. Later in the day the news sources were including the data concerning the percentages of students who took the exam. Evidently, the methodology was being explained to the news outlets.
There are always multiple possible interpretations possible when selection differences exists. Consider this as a possibility – if you included the scores from many students not intending to go to college who also were forced to take this exam in which they have no real stake wouldn’t you kind of expect the group average to be lower.
There is a second version of this story one sometimes hears. ND has SAT scores that are far above the national average. For example, in 2010 the national average math score was 516, North Dakota students scored an average of 594. Of course, only 265 ND students took the exam and these students were likely interested in out of state schools which required the SAT.
Class dismissed.