Several callers challenged a number of Michelle Rhee’s lies, distortions, and omissions today on the Kojo Nnamdi show on WAMU-FM.
One was a professor from Rutgers who pointed out that 70% of the time when a student erases an answer and bubbles in another one , the new answer is wrong.
Thus, only about 30% of the time is the direction of the erasure “wrong-to-right”.
If a student does this 12 times, I would conclude that the probability of that happening is about 0.3 raised to the 12th power, which my calculator says is about 0.0000005, or about one in two million.
But this didn’t happen to one student. It happened on the average so often that HUNDREDS of DCPS classrooms were flagged by the testing company — and the DCPS administration stalled and whitewashed it all.
Rhee Challenged on the Kojo Nnamdi Show
The URI to TrackBack this entry is: https://gfbrandenburg.wordpress.com/2011/04/04/rhee-challenged-on-the-kojo-nnamdi-show/trackback/
GFBrandenburg's Blog 
Rhee is extremely glib. Gotta hand it to her.
In one part of the interview, where she was challenged about the fact that the entire “value added” measurement method is a “black box” whose workings are unknown to anybody outside a tiny handful of economists, her response was to stonewall and to try to pit teachers against their union. She said that privacy laws prohibited her from disclosing any of that information, and that the only way that a union official could see any of the details would be if a teacher chose to share his/her scores with that official.
Of course that’s all lies. The workings of the VAM are purposely obscure and have been withheld from public view for nearly two years now. And looking at any one teacher’s alleged VAM score sheet will enlighten exactly no one as to the inner workings of that black box.
LikeLike
Looking forward to hearing the interview.
RE erasures: I had this explained to me back around 1980, by a very high figure in one of the major testing companies.
Even back then, he told me that testing companies would scan answer sheets for unusual erasure patterns, though the school system had to pay a fee. As I recall, he told me they would scan for an unusual number of erasures, and for an unusual percentage of wrong-to-right.
That is: They knew what percentage of changes would be wrong-to-right in the normal case, and they would flag the outliers, much as the caller apparently said. This stuff goes back a long ways.
LikeLike
Yeah, Rhee is the queen of glib. She’s great at making oppotunnities to fight her way out of the box she’s put herself into and gets a huge thrill when she succeeds.
How does this help the children?
LikeLike
It doesn’t help the children, but then, was Michelle Rhee ever sincerely interested in helping the children who are in DCPS?
Personally, I think that Rhee’s agenda has always been about promoting herself first, last, and always. She’s been very skillfull at attracting powerful, wealthy patrons and she’s been very willing to be their spokesperson.
LikeLike
You are absolutely correct.
LikeLike
First off, I’m no fan of Rhee. However, the Rutgers’ Prof’s assertion is very different from the literature that I’ve ran across. That is, changed answers are more likely to be correct (i.e., changed from incorrect to correct). At least this is what much of the literature suggests about college students. If there is a separate literature with grade school kids, I’d be interested.
From Kruger et al. (2005):
The majority of answer changes are from incorrect to correct, and most people who change their answers usually improve their test scores (Archer & Pippert, 1962; Bath, 1967; Clark, 1962; Copeland, 1972; Crocker & Benson, 1980; Davis, 1975; Foote & Belinky, 1972; Jarrett, 1948; Johnston, 1975; Lamson, 1935; Lehman, 1928; Lowe & Crawford, 1929; Lynch & Smith, 1975; Mathews, 1929; Pascale, 1974; Range, Anderson, & Wesley, 1982; Reile & Briggs, 1952; Reiling & Taylor, 1972; Schwarz, McMorris, & DeMers, 1991; Sitton, Adams, & Anderson, 1980; Smith et al., 1979; Vidler, 1980; Vispoel, 1998). This is true regardless of whether the test is multiple-choice or true-false, achievement or aptitude, timed or un-timed, computer or pencil-and-paper. In fact, the evidence so strongly counters the belief and strategy that one comprehensive review found that in not one of 33 studies were test-takers hurt, on
average, by changing their answers (Benjamin, Cavell, & Shallenberger, 1984).
LikeLike
You have a point. The caller did say that this was an unscientific observation of hers, and what I see is that about 58% of changed answers are correct.
Nonetheless, if one changes answers 12 times, the likelihood of all of them being correct is. now about 14 in ten thousand.
LikeLike