by Richard P. Phelps
The U.S. Government Accountability Office (GAO), a research agency of the U.S. Congress, continues its foray into the field of standardized testing. It started at least as far back as 1993 with a report I wrote on the extent and cost of all systemwide testing in the public schools. Many studies related to school assessment have been completed since, for example, in 1998, 2006, 2006, and 2009.
In the wake of educator cheating scandals in Atlanta, Washington, DC, and elsewhere, the GAO has recently turned its attention to test security (a.k.a. test integrity). For a year or so, they have hosted a web site with the fetching title "Potential State Testing Improprieties":
"â¦for the purpose of gathering information on fraudulent behavior in state-administered standardized tests. The information submitted here will be used as part of an ongoing GAO investigation into cheating by school officials nationwide, and will be referred to the appropriate State Educational Agency, the U.S. Department of Education, or other agencies, as appropriate.
"Any information provided through this website will be encrypted through our secure server and handled only by authorized staff. GAO will not release any individually identifiable information provided through this website unless compelled by law or required to do so by Congress. Anonymous reports are also welcome. However, providing GAO with as much information as possible allows us to ensure that our investigation is as thorough and efficient as possible. Providing contact information is particularly important to enable clarification or requests for additional information regarding submitted reports."
I encourage anyone with relevant information to participate, though I would be more encouraging than is the GAO about submitting the information anonymously. In some states, the "State Educational Agency" to which your personal information will be submitted is, indeed, independent, law-abiding, and interested in rooting out corruption; in others, it either does not care much about the issue or itself is an integral part of the corruption.
It would be far better if the information were not submitted to any "educational agencies" but rather to state and federal auditors and attorney generals. The problem with education agencies is that they have gotten too comfortable with their own, separate world, with its own elections, funding sources, governance structures, rules and regulations, and ethical code that places a higher priority on the perceived needs of educators than others.
This past week, the GAO released another report with a typically understated title, "K-12 Education: States' Test Security Policies and Procedures Varied". Among its findings:
"All states reported including at least 50 percent of the leading practices in test security into their policies and procedures. However, states varied in the extent to which they incorporated certain categories of leading practices. For example, 22 states reported having all of the leading practices for security training, but four states had none of the practices in this category. Despite having reported that they have various leading practices in place to mitigate testing irregularities and prevent cheating, many states reported feeling vulnerable to cheating at some point during the testing process."
Does one feel better or worse about test security after reading this passage? Is knowing that states are getting their test security policies and procedures at least half right reassuring? Would you trust your life's savings to a bank that assured you of including at least half of the leading practices in bank security in their policies and procedures?
Though the low percentage may disappoint, I find another aspect of the GAO study more worrisome: it's entirely about plans and policies and not any actual behavior. In this, the GAO takes its cue from the two associations whose test security checklist it employed in its study: the Council of Chief State School Officers (CCSSO) and the Association of Test Publishers (ATP). Their checklist comprised the "leading practices" to which the GAO refers.
Peruse the list of leading practices, included in the GAO report's enclosures (starting on p.40) and one may be surprised at how ethereal they are. Schools should have "Procedures to Keep Testing Facilities Secure", "Rules for Storage of Testing Materials", "Procedures to Prevent Potential Security Breaches", and even "Ethical Practices". There's no mention of what such rules, procedures, and practices might look like; local schools and districts are free to interpret them as they wish, presuming they even know how. Moreover, there's no mention of any actual implementation of any of the rules, procedures, and practices; in Ed-speak, test security is about having a test security plan, not actually securing tests.
In a footnote (p.2) the GAO admits "Our survey did not examine state or local implementation of these test security policies." Even if the GAO had tried to examine state or local implementation, though, what could it have found?
A "leading practice" in the terms of the CCSSO and ATP is not a "practice" at all; it is a plan for practice. That is, it is not about behavior or action, it is about a plan for behavior or action. And even the character of the plan is left to the discretion of the local school or district. Any local school or district with a test security plan in its files can claim that it is following leading practices. As model test security plans are routinely provided by test developers as part of their contract, every local school or district can be a leading test security practitioner by default.
They need not do anything to secure their tests to be a leading test security practitioner.
Borrowing a phrase so often used by the GAO in its review of other government agencies' work, the GAO "could do more" to study the issue of test security.