|Read a Post for The Expansion of Federal Power and the Politics of Implementing the No Child Left Behind Act
|Reply to this Post|
Proficiency?--In a Test's Eye
|Posted By: Dick Schutz on December 5, 2005|
|I stated "rest results [Ouch! Damn you, Spellchecker!] are commonly reported in terms of levels of proficiency....These designations do not reference student’s expertise. They are simply segments of the normalized distribution of the scores on sets of test items."|
Well, whaddaya know. Illinois, as reported by the Chicago Tribune,
inadvertently provided evidence to support the point. Turns out the State set the cut score for 5th Grade Reading at 156. But of course "156" is an arbitrary number that the
scale inadvertently omitted. As a result, 4342 kids across the state who scored 155 were designated “non-proficient” and 29 schools across the state were designated as “failing.” This could happen to any state, so it’s no blot on Illinois.
School people are not blind to these contingencies. The practice of focusing instruction on “bubble kids”—kids scoring very near the cut score—is becoming increasingly common. Kids needing the most help—way below the bubble-- are then ignored, as are kids reasonably far enough above the bubble to avoid regression below the cut score due to test unreliability.
The Trib article attributes the flaw to a “math miscue.” The glitch rather is a fatal flaw in the NCLB legislation and in the psychometric horse it rode in on.