Raw Scores, Scaled Scores, and Louisiana’s PARCC-ish Tests
A number of Louisiana parents are insisting that the Louisiana Department of Education (LDOE) release raw score information on the PARCC-sort-of test that it administered in spring 2015 to students in grades 3 through 8.
I agree that parents have the right to know their children’s raw scores. Yet these raw scores must accompany scaled scores in order to be understood.
The reason for this is that standardized testing utilizes various forms of a test. Sometimes various forms are given in the same test administration. Other times, the forms are altered as items are changed out in order to avoid unfair advantage to test takers in future administrations.
For any standardized test, a raw score is available for each test taker. However, some versions of a test are easier than other versions. Therefore, the raw score could be biased in favor of students taking the easier test. In other words, two students– one who scores a higher score on an easier version of a test and another who takes a more difficult version– might actually be “equal” in their performance when the outcome is adjusted for differences in test difficulty.
Scaled scores account for the difficulty of tests and adjust the score to account for such differences in test form difficulty. They do so via a method called “equating.”
For more on raw scores, scaled scores, and equating, see this brief, 2011 article by ETS.
Now, let’s turn our attention for a moment to Louisiana and its PARCC-ish test.
All of the other states that purported to give PARCC tests in 2015 contracted with Pearson, the vendor of PARCC tests. Louisiana gave its PARCC-styled tests via an old contract with Data Recognition Corp (DRC).
So, for all other states that called themselves “PARCC states” in spring 2015, Pearson is handling the scoring (including scaling and equating).
DRC appears to be doing so for LDOE. (The PARCC website for Louisiana notes that Raw Scores, Scaled Scores, and Louisiana’s PARCC-ish Tests | deutsch29:
