Shortlink: http://shawnbeightol.com/blog/?p=584

I just read David Smilley’s Herald article “Miami-Dade Schools considers training on conducting evaluations” and got stuck on this: “the School Board could approve negotiations with a vendor to create a new system worth up to $2.3 million to train and certify school administrators on how to conduct classroom observations.”

So, let me get this straight – a “vendor” is going to certify that administrators can certify teachers to certify students? Hmmmmm….who is going to certify the ones (vendor) certifying the ones (administrators) certifying the ones (teachers) certifying the students? And, who is going to certify the vendors?

So, we’ll also need a VAM formula to measure the effect of the principal on the teachers…and a VAM to measure the Vendor’s effect on the Administrators…and a VAM formula to…

In brief, a formula to measure the efficacy of a formula that measures the efficacy of a formula that measures the efficacy of…you get it…its like facing 2 mirrors at each other and looking in…it’s an infinite recursion.

But really, that’s crazy talk. They don’t want to carry this out to its logical limit, they just want to create/justify a new layer of bureaucracy. Hey, their friends need to pay their mortgages for their winter and summer homes too!

Mr. Smilley’s article concludes with “In Miami-Dade, Carvalho has said the formula is “mathematically sound,”

I always knew he was a better mathematician than me…if anyone can figure out “Θ ̃_j={(N_j σ_t^2)/(N_j (σ_s^2+σ_t^2 )+σ_e^2 )} * (∑_(i=1)^(N_j) r_((j)i) )/N_j where σ_t^2 is the teacher level variance, σ_s^2 is the school level variance, σ_e^2 is the residual variance, N_j denotes the number of students in class j and the notation (j)i is used to mean that student i in class j. Equation 1 above is nothing more than the scalar representation of the commonly used matrix notation: Θ ̃=DZ^’ V^(-1) (y-Wδ) where V=ZDZ^’+Ω. and V is block-diagonal”…etc ad naseum,

Our science teacher/superintendent can!

Funny, ’cause the guys paid to research BS like the Florida VAM model have to say this in their abstract on a study of the general mathematics behind specific states’ VAM models: “We also explore the potential impact of **model misspecifications**, including **missing student covariates** and **assumptions** about the accumulation of teacher effects over time, on key inferences made from the models…student characteristics are likely to **confound estimated teacher **effects when schools serve distinctly different populations.” (http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2743034/ – got my summer reading now!).

Sounds to me like there’s a bunch of room for errors (like, um, Ferguson students are a totally “different population” than Krop or Carol City or Beach High.

Anyway, for you OCD nerd-types like me, you know you’ll be dissecting this NIH and related studies this summer to understand where Florida’s VAM falls short…as we prepare for the certain lawsuits that will abound from this idiocy.

Did we really elect these people?