inBloom, Clever, and Student Privacy: More “Big Data” Considerations
My last few posts have focused on the issue of student data collection, storage, and dissemination to “education vendors.” Meanwhile, I have also been reading about the call to separate the Common Core State Standards (CCSS) from assessments. I maintain that the spectrum of reforms are intended to be a package deal and are promoted as such by the federal government–CCSS, assessments, and data collection.
One set fits all.
In this post, I will examine some more nuances related to massive student data collection as such have appeared in the comments section of previous posts. First, let me offer readers a Race to the Top (RTTT) requirements refresher regarding its full-reform-spectrum intention.
Federal Education Reforms Are a Package Deal
The federal government’s intention is for full-spectrum reform. President Obama and US Secretary of Education Arne Duncan make their comprehensive reform push obvious via the 2009 Race to the Top (RTTT)requirements– CCSS, assessment, data collection, and teacher evaluation.
Obama and Duncan want the Package Deal. First, consider this “absolute priority” for a state’s receipt of RTTT funds (abbreviated here):
Priority 1: Absolute Priority: Comprehensive Approach to Education Reform … states must… address all of the four reform areas specified in the ARRA (American Reinvestment and Recovery Act):
(Adopting standards and assessments… Building data systems… Recruiting, developing, rewarding, and retaining effective teachers and principals… and Turning around our lowest-achieving schools)
As well as State Success Factors Criteria:
(Articulating State’s education reform agenda and LEAs’ [local education agency’s] participation in it; Building strong statewide capacity to implement, scale up, and sustain proposed plans; Demonstrating significant progress in raising achievement and closing gaps).
As far as the insufficient-fund-dangling federal government is concerned, the “global success” of American public education rests in pervasive, obtrusive standardization. As for the states: They need to modify their laws in order to make RTTT mandates happen if they expect to (maybe) be rewarded funding that cannot possibly keep this standardization wrecking ball in motion:
Eligibility Requirements …
At the time the State submits its application, there must not be any legal, statutory, or regulatory barriers at the State level to linking data on student achievement (as defined in this notice) or student growth (as defined in this notice) to teachers and principals for the purpose of teacher and principal evaluation. [Emphasis added.]
Thus, the federal government requires that the states not interfere with the federal desire to use student test scores to measure teacher “effectiveness.” It sounds good, so let’s go for it.
Particularly since 1994 and Clinton’s Goals 2000, the federal government has “conditioned federal funding on state adherence to federal priorities.” This practice has only intensified with No Child Left Behind (NCLB) and RTTT.
Consequently, no interfering with collecting the data– which brings me to the primary point of this post.
Federal Education Reforms Happen Via Fiscal Lure
In my recent writings on student data mining (here and here and here), I have had commentary regarding the fact that the federal government has been collecting data for decades. This is true; however, according to the US Department of Education (USDOE), “there are no unfunded federal education ‘mandates.'” Thus, USDOE is technically correct when it reports that state involvement in such programs as RTTT is “voluntary.” Never mind that they did so to escape NCLB, and never mind that in this day of reform, one or two signatures is all that it takes to commit an entire state to the likes of inBloom (secretly contracted by Louisiana State Superintendent John White– one signature) or CCSS (only requiring two signatures).
inBloom vs. Clever
Another comment on my inBloom posts involves the fact that inBloom is not alone, that there are other data “cloud” services. This is true. One notable such company is Clever, a San Francisco-based company started in mid-2012 that serves 15,000 schools in Los Angeles, Nashville, and Houston.
Nevertheless, Clever and inBloom differ in some key ways. First of all, Clever is not the direct creation of those funding and promoting CCSS and the rest of the RTTT reform spectrum. As a result, districts choosing Clever are free of the straitjacket smothering that accompanies The Reforms as such are being pushed onto states by RTTT.
I am not advocating that issues of student privacy should be ignored since Clever is “outside of” the RTTT stronghold. However, at this point, Clever has not been sucked into RTTT, and Clever is not (yet?) actively advertising that its vendors are aligning their education products to CCSS standards. This makes me think that the districts storing data with Clever have control over whether the student data they choose to store is de-identified (i.e., it cannot be traced to specific students). I do not know this for sure; I only know that the absence of certain fingers on Clever (e.g., Gates, the National Governors Association, the Council of Chief State School Officers, and Duncan, to name a few) makes student privacy more likely since districts choosing Clever are not under “reformer coercion.”
A converse key issue in the Clever-inBloom comparison is the fact that inBloom was created expressly to “unite reform.” InBloom is the result of the Gates-Carnegie-funded Shared Learning Collaborative. From its inception, inBloom was intended to offer a venue for CCSS-aligned educational products. Thus, the vendors are tailoring their products to produce CCSS-aligned curriculum that “states” might “voluntarily” adopt as CCSS-aligned assessments loom in their future.
(This interesting article on student information systems [SIS] notes that in order for these data sharing systems to work, there must be “an openness among vendors”–for one company to allow another to build “product upon product.” The author notes that Pearson is willing to participate in this process. It just so happens that Pearson is connected to both the PARCC and Smarter Balanced testing consortia. Of course, in Los Angeles, Pearson is already cashing in on its CCSS software for Apple iPads.)
Another distinction between inBloom and data storing companies such as Clever is that inBloom is a nonprofit. As such, inBloom does not pay taxes, and those donating to “the charity” that is inBloom are able to deduct their donations on their taxes. Furthermore, those with a key interest (usually fiscal) in inBloom direct its board. Current inBloom directors hail from institutions actively promoting privatization; Gates, Carnegie, and CCSSO.
Envision the future of inBloom, secured by pro-privatizer, philanthropic funding. Envision the USDOE offering states grant money to “voluntarily” use inBloom for their data storage in order to centralize student data. (If this sounds far-fetched, realize that USDOE enticed the centralizing of math and ELA [English Language Arts] standards by dangling money in front of “states” for the “voluntary” adoption of standards that had yet to be written.)
This is one way that inBloom can become a guaranteed monopoly– USDOE enticement of states via financial lure: States, just agree to store data on your students with inBloom as part of some new Race to Somewhere That Top-notch American Private Schools Refuse to Go, and USDOE will give the chance (nothing definite, mind you) for your state to receive some Federal Dough.
The inBloom “Sandbox”
Yet another comment to my inBloom posts referred to inBloom’s offers its infrastructure free to the public for download. How is it that inBloom could possibly become a monopoly if it gives itself away? (I stand corrected that it is not one yet, but I maintain that its backers position inBloom for unrivaled privatization success.)
As information available on its site notes, inBloom is not giving away its infrastructure; it is offering a demo in hopes to encourage the development of inBloom-compatible applications. As such, the “sandbox” offers education software developers the opportunity to create products from which inBloom might make a profit:
The sandbox environment is a full replica of the production environment with fixture data, designed for developing and testing application prototypes. Each sandbox account is a logically distinct tenancy that mimics the production Secure Data Service while being completely isolated from any live data. Sandbox users are typically vendors and individual developers wishing to try new ideas or develop systems without impacting existing services.
Sandbox environments are not for use with real data containing Personally Identifiable Information (PII) and user may experience different performance levels than official production instances.
Rest assured, inBloom is not undercutting its existence by offering the public this “sandbox”; inBloom is securing its place in public education standardization. (See update below for a reader comment on the sandbox situation.)
The Potential for Holding Data Hostage
One of the issues I envision for this profound, ubiquitous student data storage is the problem of fierce competition among “cloud” companies. I envision a potential constant changing of vendors similar to that brought about by competition among phone companies or cable/dish television providers.
Granted, changing data “clouds” is not as easily accomplished for states as is changing a cable TV provider for an individual. Then there’s the question of whether those operating a data “cloud” might refuse to release data (or might shut down a state’s ability to access data) if a state attempts to remove data from one “cloud” provider based upon enticement by a more attractive provider.
And what of the security of the data as it is transferred from one “cloud” provider to another? Given the absence of a regulatory agency to oversee the quality of “cloud” providers, are not states vulnerable to those who would create “cloud” companies with the intent to steal data and sell it for profit despite any agreement to keep the data secure?
This brings me to my final point of this post:
We Need a Regulatory Agency to Oversee Student Data Storage, and We Need Legislation to Mandate De-identification of Student Data
I am against the unprecedented student data collection required as part of the package of reforms fiercely promoted by Duncan, NGA, CCSSO, Gates, and others exempt from experiencing the impacts of their so-called education reform impositions. I have been told to “suck it up” and accept that the reforms will happen, not the least of which is massive data collection.
Just give it up.
Those pushing reforms in the name of “accountability” need themselves to be held accountable. Thus, if they want student data, they need to agree to first establish a regulatory agency (one untouched by philanthropic or USDOE funding) to oversee would-be data storage organizations.
Establishment of such accountability should have preceded the demand (call it “voluntary” if you want, Arne, but exit from involuntary NCLB into RTTT is hardly freedom of choice) that states collect data intended to be tied to CCSS, its assessments, and ultimately connected to teacher (and school, and district) livelihood.
The effectiveness of a student data storage and usage regulatory agency hinges upon legislation protecting students from unnecessary (this must be carefully defined) release of data markers enabling student identification. As a researcher, I am ethically responsible for making the data upon which I base a study available to other researchers for study replication (a lost art in this atmosphere of data-hiding-and-twisting corporate reform). I am also responsible for de-identifying my data before making it available to third parties. Study participant privacy matters, as does student privacy, regardless of what Duncan has done by way of FERPA castration.
To states that have taken the Duncan cash and have weakened their states’ data collection and usage/student (and teacher) privacy laws for the purpose of “testing” teachers based upon student data: Give the RTTT money back. Reinstate the student/teacher privacy laws.
To states that did not have strong laws in place to protect students and teachers from the heightened, potential exploitation fostered by the current, profit-driven atmosphere fostered by corporate reform: Enact data privacy laws to protect your constituents.
Washington State has a student privacy bill ready for the legislative session that begins on January 13, 2014. New York is raising support for a parent “opt in” student data collection bill. New York is the last “fully committed” state left in inBloom’s plan to pilot its “cloud.” Louisiana is also working on student privacy legislation. These are only examples.
It is no surprise that in all three states the student privacy focus was spurred by the RTTT reform package.
Call it RTTT reflux.
I’m sure that more will, uh, come up.
UPDATE 01-12-14 on “sandbox” implications:
From a commenter:
Your points about open source are a bit mixed up though. The “sandbox” is just a demo, yes. But the code to the data store is available on github, and all its additional components are also available as free/open source software, so anyone with some Java and other relevant skills can run the same server as inBloom, on the same underlying cloud service if they’re using Amazon as reported.
The software isn’t really the issue here — it is the organization. If individual states or districts hired their own sys admins to run their own inBloom servers on their own hardware, it wouldn’t be controversial.
Put another way, inBloom the organization could be completely destroyed, and Pearson could quiety set up their own inBloom servers and continue the effort internally.
What they’re proposing to do is create the equivalent of the credit rating agencies for educational data, and we all know how powerful and unaccountable those businesses are, and how far their influence has gone beyond what people imagined at their creation (i.e., needing a good credit score to get a job). What software they’re running, or if I can set up a copy of it at home, isn’t really the issue.