A couple of months ago I warned about the dangers of having government gather and publish growing reams of information in the name of making education better. Sure, it sounds great – help people get as informed as possible! – but the dangers are legion. You can read about several such pitfalls in that old post. You can also get a sense of the great wealth of data already out there in this op-ed. What I haven’t discussed – and what might concern many Americans more than anything else – is the threat that massive data collection poses to our privacy.

Articles over the last week or so have started to draw significant attention to the growing education-information complex and its connection to long-standing efforts – especially federal – to accumulate information on Americans from birth to boardroom. Gaining particular traction has been a story about how student data collected in New York could be sold to companies or other entities outside of school districts. Even more concerning is a story by Joy Pullmann in the Orange County Register about lots of data collection and mining that is either already happening or under consideration nationwide.

What’s especially troubling to some people, including Pullmann, is that not only is there ever-growing centralization of curricula such as the federally backed Common Core, as well as centralized testing of knowledge, but there are also moves to assess students’ “affect” that could include wiring them to “facial expression” cameras and “skin conductance sensors.” Contemplating such things, it’s hard not to conjure up images of A Clockwork Orange.

When you read the federal report that proposes using “affective computing methods” such as skin sensors, it doesn’t appear that the authors have nefarious, big-brother intentions. The object of the report is to examine how students’ “grit” and perseverance can be improved, and that is a reasonable goal. Similarly, furnishing information about the academic status of incoming freshmen at a college, the amount they learn while in school, and how well they fare after graduation, is driven by good intentions.

But we must never feel content with good intentions. We must care primarily about the effects of the policies stemming from our golden goals, and as I’ve written previously, there are likely big, negative, immediate effects that would go with empowering more government data collection. There are also potentially even worse long-term consequences, including that government would begin to try to adjust students’ feelings and attitudes if doing so might produce better test scores or some other, politically determined, outcome. Indeed, such affect-engineering arguably already takes place with huge increases in ADD and ADHD diagnoses that lead to personality-altering drug-taking.

It’s easy – and almost always innocent – to say that we need more information so that we can make things work better. But with that comes very big potential dangers we must never ignore.

Cross-posted at seethruedu.com

Though there are currently more students participating in scholarship tax credit (STC) programs than voucher programs nationwide (about 151,000 to 104,000), the former have not received nearly as much attention as the latter. That has begun to change in recent years as growth in the number of STC programs has outpaced growth in voucher programs.

Over the past week, I have enjoyed engaging in a spirited debate over STC programs with Professor Kevin Welner of the University of Colorado at Boulder. The debate was sparked by Valerie Strauss’ blog post at the Washington Post, which contained several significant errors that I addressed here. Welner then responded at Strauss’ blog and we continued to spar here and here. It is my sincere hope that readers who have followed the debate have found it illuminating.

Though I suspect that Welner might not share my aspiration for universal educational choice, we have a least found common ground in the belief that, given limited resources, such programs should first aid those most in need. I also agree that our three primary areas of contention are: 1) the differences between STC programs and vouchers and their significance; 2) the fiscal impact of STC programs; and 3) who receives tax-credit scholarships. I will address Welner’s latest arguments on these matters below.

First, however, I must make two important corrections to Welner’s last post. In explaining why he did not provide context for some of his remarks, Welner wrote: “Much of this change happened in the aftermath of the 2010 midterm elections, when Republicans swept into state offices in very large numbers.” Actually, only five out of the fourteen STC programs (in Louisiana, New Hampshire, Oklahoma, Virginia and one of the two in Pennsylvania) were adopted in the wake of the 2010 midterm elections. Moreover, there was strong support among Democrats in two of those states. Pennsylvania’s 2012 STC legislation passed with the support of 15 of 20 Democrats in the Senate and unanimous support in the House. In Louisiana, the legislation passed with the support of 11 of 15 Senate Democrats (32-7 total) and 32 of 45 House Democrats (66-37 total).

Scholarship Tax Credits vs. Vouchers

Welner wonders why I did not use the term he invented to describe scholarship tax credit programs. “Neovouchers” is a confusing term that appears nowhere in any of the fourteen STC laws. It also blurs the distinctions between STCs and vouchers, which I have described previously. I have likewise avoided the term “opportunity scholarships” because it is essentially meaningless as well. The terms “scholarship tax credits” or “education tax credits” accurately describe a program in which individuals or corporations receive tax credits for donating to scholarship organizations that fund low- and middle-income students attending nonpublic schools. I don’t begrudge Welner for using the term that shares a name with his book, but I also don’t see why he should expect that others should adopt it.

In my previous posts, I argued that these two policies have similar ends but very different means and therefore should be called by different names. I then explained how the means are different, particularly their funding (public vs. private money) and administration (government-run/centralized/uniform vs. privately-run/decentralized/diverse). Welner then responds, essentially, “Yes, but their ends are nearly identical!” I would suggest that he misses the point.

Welner also takes issue with the examples I gave of courts that decided the question of whether tax credits constitute public or private money. Welner noted correctly that some of those cases did not pertain directly to scholarship tax credit programs. What he misses is that this fact strengthens my point. State courts have ruled that tax credits do not constitute “public funds” both with regard to STC programs and other forms of tax credits. This consistency shows that STC programs are not merely a legal loophole or “money laundering”, as Welner called it. The freedom of citizens to direct their own money makes such tax credit programs qualitatively different in policy terms and this difference is reflected in the law, not arbitrarily invented by it.

Credible Evidence of Savings

Welner points out that I overinterpreted his statement that he would not be surprised if Florida’s STC program generates savings. Instead, he holds that the available evidence does not support that conclusion. He argues that we do not have all the data necessary for a conclusive determination so he throws up his hands. In fact, there is credible evidence of savings.

The best available estimate of any STC program’s fiscal impact is from Florida’s Office of Program Policy Analysis and Government Accountability (OPPAGA). This is important since Florida’s STC program is the least likely candidate for realizing savings (with the possible exception of Georgia’s). Florida offers the maximum possible tax credit (100%) whereas programs in seven of the other ten states offer only partial credits, as low as 50% in Indiana and Oklahoma. Florida has the largest average scholarship size and the highest ratio of scholarship size to average public school operating per pupil expenditures, as shown in the table below. (Note that the National Center for Education Statistics’ calculation of total per pupil expenditures excludes unfunded pension liabilities. Moreover, low-income students generally cost the state more money than average to educate.)

[This chart excludes Arizona’s STC program for special needs students and the STC programs in Louisiana, New Hampshire, Oklahoma, Virginia and Pennsylvania’s corporate program, which were only recently launched or have yet to launch. Table includes the most recent data available in each category.]

Now Welner is certainly correct that savings depend on the ratio of switchers to stayers, but the data I’ve provided thus far indicates that the percentage of switchers does not have to be very high to realize savings in most states. Welner was rightly skeptical of OPPAGA’s 2008 report, which made an educated guess that 90% of scholarship recipients were switchers. However, OPPAGA’s 2010 report and 2012 revenue estimating conference relied on U.S. Census data and found that their previous estimate of switchers had been too low, since 94.6% of scholarship-eligible low-income students were attending public schools in the year before the STC program took effect. As Jon East explained in RedefinED, “The estimating conference went even further, combining American Community Survey data from 2005-09 with private school enrollment data to make projections about the actual number of low-income students enrolled in each grade level in private schools in 2012.” The more recent report projected savings of $57.9 million for Florida in 2012-13.

Welner is also correct that the analysis of the total fiscal impact of STC programs should not stop there. States that offer less than Florida’s 100% tax credit should also account for the impact of the deduction of non-credit eligible portion of the donation, as well as the caps on deductions. A complete fiscal analysis would also have to include other government programs or tax credits that are available in a given state. I agree with Welner that in most states, we need more data. However, the evidence of savings in Florida is strong, even accounting for Welner’s caveats. And if there are savings in the least likeliest of states, then there are likely savings elsewhere.

Clear Benefit to Low-Income Families

In my previous posts, I criticized Strauss for claiming that low-income families do not benefit from tax-credit scholarships. Welner admits that STC programs “provide financial assistance to many lower-income families” but says that he “didn’t read [Strauss’] statement to be saying that zero low-income families receive neovouchers. ” Once again, Strauss correctly noted that tax-credit scholarships do not cover the full cost of tuition, then incorrectly concluded: “Poor families can’t make up the difference. Guess who can.” That’s a fairly unambiguous statement. Strauss didn’t even qualify her claim by referring to “most” or “some” low-income families, let alone provide any evidence to support her claim. If she wants to be taken seriously as a responsible commentator, she should correct the record.

Likewise, Strauss has not yet rescinded her fallacious charge that STC programs are “welfare for the rich” because the donors somehow benefit from the tax credits. As I have demonstrated, the donors break even at most. Even Welner abandoned that line of argument in his latest post. Again, Strauss has a duty to correct the record.

In his latest post, Welner conceded that all of the STC programs are means-tested but for Georgia’s and one of Arizona’s two programs. However, Welner expressed skepticism about the organization that issued the study showing that two-thirds of scholarship recipients in Arizona fall under 185% of the federal poverty line. He also noted correctly that the income thresholds in some states allow some middle-income families to qualify as well. That said, it is unclear why he ignored the evidence I provided from state governments showing that the average income of scholarship recipients is far below the means-testing thresholds. For example, the average income of recipient families in Pennsylvania was only $29,000, just under half of the state’s income threshold at the time. Welner has not explained why we should assume that recipients in other states look significantly different, especially when there is evidence of similar patterns.

Welner calls for more a more comprehensive state-level reporting system. I am sympathetic to this suggestion, though I believe that states should proceed with caution. Scholarship organizations are already more regulated than ordinary nonprofits, like the Salvation Army or Red Cross. While regulations vary by state, STC programs generally have more stringent accounting standards, reporting requirements, and some states even require background checks for employees. Every STC program requires that scholarship organizations spend no more than 10% on administrative costs, the exceptions being Florida’s 3% maximum and Pennsylvania’s unnecessarily high 20% maximum. (It’s important to note that a government study found that 62% of Pennsylvania scholarship organizations disbursed 100% of their collected funds while only 5% used the maximum administrative expenses.)

Our education system should empower families to choose the education that best meets their kids’ individual needs. Scholarship tax credit programs move our education system toward that goal. As with all government programs, we should constantly reassess whether STC programs are achieving their desired ends and make any necessary changes. I would like to thank Professor Welner for taking the time to discuss this important matter.

[Update: An earlier version of this post incorrectly labeled total per pupil expenditures as operating per pupil expenditures.]