George Will's Urban Legend
In his column of Aug. 26, 1993, the syndicated columnist George F. Will wrote, "Nationally, about half of all urban public school teachers with school-age children send their children to private schools.'' He repeated the statistic a few days later on "This Week With David Brinkley.'' In so doing, Mr. Will perpetrated a new urban legend. "Urban legend'' is the phrase affixed to widely circulated, sometimes widely believed, contemporary stories that, on close examination, turn out to be fantasy. Crocodiles in the sewers of New York City is perhaps the most commonly told of the genre. The subterranean city under Chicago started by survivors of the Great Fire is another. Mr. Will's statistic is yet another pure figment.
If Mr. Will's number were true, it would certainly be a stunner: The 1984 Metropolitan Life Survey of Teachers found that, nationally, only 14 percent of public school teachers with school-age children sent them to private schools. Since 12 percent of the nation's schoolchildren are in private schools, the Met Life survey reveals that public school teachers put their faith in public education almost as much as does the general public. In fact, a greater percentage of public school teachers likely uses public schools than do other families in similar socioeconomic circumstances.
Where did Mr. Will get his tally? A call to his office led first to a Litigation Backgrounder, "School Choice Cases,'' by Clint Bolick at the Institute for Justice in Washington. The I.F.J.'s director of communications, John Kramer, advised that Mr. Bolick took his figure from Liberating Schools: Education in the Inner City, by David Boaz of the Cato Institute. Mr. Boaz, in turn, said that he had gotten some of his figures from Mr. Bolick, revealing a chummy circularity of sources and mutual citations. The prime source of Mr. Boaz's numbers, though, had been a 1986 paper by Denis Doyle and Terry Hartle, both then of the American Enterprise Institute.
Commentators of all political stripes choose statistics that fit their needs, but it is worth noting that with the exception of Mr. Hartle, most of the named persons are conservative school critics and ardent advocates of school choice. Several mentioned the figure 46 percent which, conveniently rounded upward, would yield Mr. Will's national figure of "about half.''
Mr. Doyle thought Mr. Will had exaggerated: "A bit of an overstatement,'' was his comment on the 50 percent figure. A bit of an overstatement? The kind of exaggeration contained in Mr. Will's remark is the kind of exaggeration that turned Pinocchio's nose into a jousting lance.
Even if the number were accurate, to toss it off, as Mr. Will did, as an established, unchallengeable truth was, at best, unwise. Mr. Doyle and Mr. Hartle knew their work was not definitive. They titled their paper "Where Public School Teachers Send Their Children to School: A Preliminary Analysis" (emphasis added). In the body of their work, they were appropriately modest about their methodology and their findings: "The evidence in this paper is exploratory,'' perhaps sufficient reason to explain why the paper was never published.
Mr. Doyle and Mr. Hartle did not directly survey teachers in various cities, but, rather, worked from 1980 Census data which contained information on residence, occupation, children, and the schools the children attended. Their data represented only 12 states, although these states did represent all regions of the nation. Their research can be considered a reasonable first step in determining an answer to the question, but it was hardly the last word, as they were well aware.
Mr. Doyle and Mr. Hartle found that in no state over all, nor in the urban areas of those states, did anywhere near 50 percent of public school teachers with school-age children pack their kids off to private schools. In the 12 states analyzed, an average of 16 percent of all public school teachers' children were enrolled in private schools, compared with 21 percent of urban school teachers' children. The statewide rates ranged from a low of 11 percent in New Mexico to a high of 23 percent in Louisiana, a state notorious for underfunding public education. For urban areas, rates ranged from a low of 15 percent in Arizona and North Carolina to a high of 35 percent for Louisiana. Nationally, 12 percent of all children are enrolled in private schools, while 17 percent of urban children are somewhere other than in public schools. For some states, the difference between the rate at which teachers enroll their children in private schools and the rate for parents at large is as small as 1 percent; for no state is it larger than 7 percent. The national difference between teachers and parents generally is only 4 percent.
Because choice and vouchers are much on the minds of Californians these days in view of the upcoming vote on Proposition 174, let the record show that 19 percent of California public school teachers' children attended private schools, while 21 percent of California urban teachers' children did so. For California parents as a whole, the figures are 14 percent and 18 percent, respectively.
Still, one might argue that teachers reject public schools because they do enroll their children in private schools at slightly higher rates than do parents in general. However, it is not valid to compare teachers against parents as a whole.
Consider that 80 percent of teachers are women and that the Met Life survey found that 75 percent of teachers are married. This means that a large percentage of them are likely not the sole wage earner in the family. Nationally, the average teacher makes about $35,000 per year, not a princely sum by any means and certainly meager given the importance of what they do, but well above the nation's average income. If we assume that the teacher's spouse also makes $35,000 a year, this yields an annual household income of $70,000. A 1992 study by James Coleman of the University of Chicago found that the average family income of parents with children in private schools was $70,000. Thus, the typical family with a teacher in it has an average income equal to the typical private school family and an income more than double that of the typical American family.
Affluent families are more likely to send their children to private schools than are middle-income or poor families. While the superiority of private versus public schools is very much in dispute these days (recent analyses of data from the National Assessment of Educational Progress show tiny differences in spite of the advantages enjoyed by private schools), many parents certainly believe their children will get a better academic shake in private schools. Thus, in the American tradition of "giving my child every advantage possible,'' a number opt to pay for their children's education. Naturally, the often pricey tuition of such schools makes the option more readily available to families of means.
To determine if a disproportionately large percentage of teachers are sending their children to private schools, we need to compare the percentage of teachers with children in private schools with the percentage from families in the same income bracket as teacher households. This has never been done.
So where did Mr. Will and the others get their mythical 50 percent statistic? From Mr. Doyle and Mr. Hartle. And they got it wrong. In the preface to their own analysis, Mr. Doyle and Mr. Hartle claimed that the "Chicago Reporter found that 46 percent of public school teachers who live in the city [of Chicago] send their children to private schools, compared with 22 percent of all Chicagoans.''
But the Chicago Reporter said no such thing. The May 1984 Chicago Reporter, analyzing 1980 Census data with the aid of the University of Illinois, actually said that "38 percent of teachers who work for the Chicago Board of Education and live in the city send their children to private schools.'' Thirty-eight percent, not 46 percent. As with the other figures used earlier in this article, the 46 percent figure cited by Mr. Doyle and Mr. Hartle is the percentage of public school teachers' children in private schools.
This is an important point. The Doyle-Hartle analyses indicate 16 percent of public school teachers' children and 21 percent of urban public school teachers' children in private schools. But, unless one makes the unlikely assumption that each teacher has only one child, the proportion of teachers with children in private schools will be smaller than the proportion of teachers' children in private schools, although how much smaller cannot be reckoned from the Doyle-Hartle paper. Thus, the 16 percent and 21 percent figures given above actually overstate the percent of public school teachers with children in private schools.
So, the figure cited by George Will as a national statistic is really only the usual figure given for Chicago schools. This usual figure for Chicago schools is erroneously high and the stuff of a new urban legend. The figure for many other cities is lower still. Given the terrible urban school conditions described in the annual "Bracey Report on the Condition of Public Education'' (Phi Delta Kappan, October 1991, October 1992, and October 1993) and by Jonathan Kozol in Savage Inequalities, we can certainly ponder why the figure is not 50 percent or higher. Meanwhile, next time you're in Manhattan, watch out for those alligators. And be wary of the undocumented statistics put forth by pundits like George F. Will.
Vol. 13, Issue 04, Pages 29-30Published in Print: September 29, 1993, as George Will's Urban Legend