|Read a Post for Using Diagnostic Classroom Assessment: One Question at a Time|
|Reply to this Post|
Why Add the Gobbledegook?
|Posted By: Dick Schutz on June 20, 2007|
|This all makes eminent sense. But these considerations should characterize all VALID tests of instructional accomplishments and student achievements. If the test items individually and collectively don't provide useful information to teachers, how can one contend that the test is "valid."|
There is no reason to relax reliability or any other psychometric requirements. In fact, reliability will be enhanced because kids will be picking answers they "know" are right (even though they are misinformed) rather than guessing among plausible "foils"
Apply Occam's razor. Strip all of excess terminology. "Formative" "Assessment" and "Diagnostic" all carry excess baggage that is irrelevant and unnecessary. (No need to go into the dirty linen in that baggage here, unless anyone questions that these are very squishy terms). What you're left with is a measure that is directly aligned with "standards." If a standard is "empty" and it's not feasible to write items that nail it down, that too is useful information.
Such tests can be "normed" per standard practice. The fundamental difference is that the tests are not developed per Item Response Theory which has been bastardized in stretching from the assumption of a univariate, homogeneous "latent trait"--which instructional accomplishments and student achievements are NOT.
The prime implications of the article are not for teachers. Teachers don't have the time or the background to construct such tests, and primitive "hunting and gathering" by individual teachers makes no sense. The implications are for the psychometric profession and the testing industry. Both, however, are resistant to change, so "gains" will be difficult to achieve, and "proficiency" is unlikely to be attained in the foreseeable future.