Consider a household with four family members, two adults and two small children. We could compute the average size and weight of the family, but we would end up with data points that don’t describe anyone in that family.
Education often falls into that trap. It’s natural to want to use averages to describe a huge group of individuals, but the result is often extremely misleading.
Take concerns about average test results for standardized tests like the ACT or SAT. The tests are elective (in most states), so changes in the population taking the test can affect the overall averages. In New Jersey, every college is now SAT optional, meaning that no students have to take the SAT. Other states require all students to take the SAT, whether they plan to attend college or not. Both changes affect the test-taking population and therefor potentially shift the average score in ways that do not represent student learning.
Put another way, if Pat was a typical SAT taker ten years ago, Pat might today get exactly the same score as a decade ago, but the cohort of people taking the test with Pat has changed enough to make the average score change. The change shows a change in the test-taking population, not in the test-taking capabilities.
Or consider average teacher salary. This is also affected by the population in the sample. If a district saw 10% of its experienced teaching staff retire in one year to be replaced with a cadre of first year teachers, the district’s average teacher salary would drop, even though each individual teacher was being paid just as well as ever.
Pennsylvania is an example of how averages can be used to obscure serious issues. David Lapp, Director of Policy Research for Research for Action, a nonprofit education research organization testified before the state’s Basic Education Funding Commission. Commenting on the state’s extreme education inequity, Lapp pointed out, “Every metric tells the same story. Pennsylvania boasts average or slightly above average achievement, either in funding, in opportunity, and in overall achievement.” But those averages conceal “nation-leading inequity. There are gaps in funding, gaps in opportunity, and gaps in achievement.”
For example, per pupil spending in Pennsylvania (20-21) ran from lows of around $13, 000 to highs around $40,000, making talk about the average of $19,667 somewhat meaningless. District size ran from under 500 students up to highs around 25,000 (Philadelphia is an outlier with over 200,000 students), so talking about the average district size of a little over 3,000 isn’t particularly useful.
Numbers can be crunched to create a picture of an average school district, attended by average students being given average opportunities and average resources resulting in average achievement. The problem with using any of those as a basis for policy discussion is that none of them exist. When you go to the doctor, you don’t want them to treat you for the average complaint of their average patient; you want them to deal with the specifics of your situation. Schools and classroom teachers deal with specifics, not averages.
Read the full article here
