THE ASTOUNDING EXTENT OF NATIONAL CAPITAL ILLITERACY
In mid-March, 2007, the Washington Post ran a two-column article under the modest headline:
High Rates said to Hurt Economy
Study Finds High Rate of Illiteracy
The major gist of the article is that "only 8 percent of DC residents with the lowest literacy skills get the remedial assistance they need, according to a new report by the State Education Agency."The article goes on to recount the report's arguments in favor of a population that can read.
And, oh yes, of the 469,000 adults (over 16 yrs old) in our nation's capital city (in 2000), some 170,700 "have trouble comprehending bus schedules, reading maps, and filling out job applications". If DC met the nationwide average illiteracy rate, over 70,000 more DC residents would be able to read the simpler articles in the city's newspapers, and hold better jobs. This analysis was prepared by a contractor funded by DC's "State Education Agency" (DCSEA), an inflated, if not misleading, adjunct of the University of DC, itself a dubiously titled institution.
While the DCSEA is driven, quite understandably, by the desire to get the funding (and provide the programs) for alleviating illiteracy, NARPAC felt the need to probe the validity of the substantiating analysis, and also to look more closely at where the city's functionally illiterate are now living and working. We conclude (not surprisingly) that the analysis has some shaky elements to it, but that this ghastly national embarrassment may actually be underestimated.
NARPAC is deeply discouraged by the level of illiteracy in the nation's capital (which the supporting reports may well underestimate), and very disappointed by the sophomoric rhetoric of DC's "State of Adult Literacy" report from its "State Education Agency)", which fails to differentiate either the consequences or the solutions by the cohort to which they apply. Here are our conclusions:
Overall, the functional illiteracy of the residents of the US national capital city is 75%, if not 100% worse than (i.e., twice) the national average. And the US itself is no prize, ranking 38th among the nearly 200 nations in the CIA's "World Fact Book". It would be difficult to find a better definition of a national disgrace.
While the report focuses on Adults Over 16, the major reason why DC's illiteracy rate is so high is that almost all DC adults score "below basic" in reading skills whether or not they finished high school. While the DCSEA dances around the aggregate issue swinging platitudes, DC's most pressing literacy issue is to teach its kids to read before they reach high school or drop out.
DC's second most important literacy problem (also ignored in the report) is its very high share of Mostly Single Parents who cannot stimulate their kids' desire to read. While the DCSEAR suggests that "children provide motivation for their parents to learn to read", we think the parents should motivate their kids by example. There must be some residual obligation for DCPS to promote reading skills for expectant moms to help break the 'Cycle of Poverty by Illiteracy'.
Oddly enough, DC's laborforce is not significantly diminished by its high illiteracy rate realtive to the US as a whole. Whether DC's less skilled workers work in or outside the city is not known, but improved literacy will improve both those workers' and their city's income and quality of life.
DC's Senior Citizens are not disproportionately illiterate, and exceed the national norm for college education. But one of the DCSEAR's most understated arguments is that "(literacy is) a stronger predictor of health status than age, income, employment status, education level, or race."
Finally, the DCSEAR neglects the potential impact on DC Voters. By probing their consultants' technical report and adding (almost surely inflated) DC voter registration numbers for each Ward, it is clear that over 40% of DC's voters are probably functionally illiterate, and this fraction may exceed 55 % in the city's worst educated Wards.
For those skeptics who like to trace quantitative analyses back to their roots, this effort is an interesting case in point. "Analgenealogically" speaking, the figure of 170,700 functionally illiterate DC residents flows from a mixed heritage:
o The Institute of Educational Sciences (IES) within the US Department of Education (DoEd) includes a National Center for Education Statistics (NCES);
o Every decade or so, the NCES produces a National Assessment of Adult Literacy (NAAL). The first survey of this century was conducted in 2003, and the subsequent report (NAAL03) has only recently been released. Its results are compared to those of the prior survey of 1992 (NAAL92);
o The NAAL surveys use sophisticated procedures to develop literary "test scores" (0 - 500) in each of three reading areas: "prose"; "documents"; and "quantitative", and assign those scores to four different "levels" of achievement: "below basic" (0-209); "basic" (210-264); "intermediate" (265-339); and "proficient"(340-500). (parenthetical examples are for "prose" capability);
o Any particular group (age, gender, race, language, etc.) within the total sample is then characterized by its average score (say, 275), and the percentage falling within each level. The "below basic" category is labeled "Level 1", and its occupants described as essentially "functionally illiterate". The differences are most stark among the races: for instance, in NAAL03 only 7% of white adults were at Level 1, but 44% of Hispanics were deemed functionally illiterate;
o Similar to the NAAL92 survey, the NAAL03 survey included over 25,000 households, distributed among some 160 counties in about ten states. (DC was not included.) Respondents were identified by educational achievement (7 categories), race (5), gender (2), disability (2), immigration status (2), prior year employment (5), occupation (4), and labor force status (3);
o It seems remarkable to NARPAC that there was no separate categorization for either parental status, parental education status, or household income, all three of which we consider fundamental to the likely educational outcome for any individual;
o These surveys were essentially national in scope and intent, and the average scores for any one variable ignored variations in others. In essence, the results, while informative in the aggregate, do not allow "synthesizing" equivalent scores for other specified population mixes.
o The ability to "work the problem backwards" was resolved (at least partially) by Stephen Reder of Portland State University in 1997 for the Office of Vocational and Adult Education within DoEd. Working with NAAL92, he derived "regression coefficients" (to us, the equivalent of "partial derivatives") for many of the individual variables (race, gender, education, etc.)
o To provide a statistical basis for its newly released State of Adult Literacy Report, DCSEA turned to consultants at the Kairos Management (with no particularly evident expertise in this particular field). They essentially applied the 1990 "Reder method" to the eight wards within DC, in order to come up with a number of functionally illiterate residents in each Ward in 2000.
o Kairos appears to have spent much of its 18 month effort developing demographic characteristics for each of DC's eight wards, as re-defined in 2000, from 1990 Census tract data. There does not appear to have been any formal requirement to assess whether or not this (only available) regression approach had limitations, or applied as well to 2003 data as to 1992 data. In essence, the Kairos estimates were derived entirely on 1990/92 data, and jiggled to match to 2000 Census population data. So DCSEA now knows with great precision that among Ward 8's 47,604 adults (over 16) seven years ago, 23,278 are functionally illiterate (i.e., "below basic"), and 41,177 are at or below "basic". Only some 6,427 Ward 8 residents (13%), then, are "intermediate" or "proficient" at reading, while 84% of Ward 3 residents are in those upper levels.
o Kairos's Technical Appendix to the State of Adult Literacy Report moves one step closer to NARPAC's major interest: what on earth are these 171,000 functionally illiterate souls doing? Through a relatively simple set of assumptions (called a "thought experiment" by Kairos, but just a "what if" excursion by NARPAC), the report concludes that 57,000 are over 65 and make up the bulk of those retired. Of the remaining 114,000 of working age, the report further concludes that 71,000 are "not in the labor force". This leaves a mere 43,000 in the workforce marginally able to read their own health or retirement benefits (if any).
Most concerned observers should find these statistics on DC's illiteracy rate mind-boggling whether or not they are precisely correct. Does it make any real difference whether the percent of functionally illiterate adults in our nation's capital city is 30%, 36% or 42%? Probably not. Any number placing Washington significantly above the national average of 21% should be a major cause for both concern and embarrassment. In fact the national average itself should be a major cause for concern and embarrassment. Not mentioned in the DCSEA report is the fact that the US as awhole does not stand anywhere near the top among the nations of the world. According to the CIA fact books, the US holds 38th place: behind most developed, West and East European, Asian and Central Asian countries. Nevertheless, here are several NARPAC questions concerning this "State of DC's Adult Literacy" effort and our own answers to them:
Is a 36% adult functional illiteracy rate plausible?
Apparently it is. NARPAC is in no position to challenges the responses to the National Adult Literacy Surveys, and provided its first summary of NAAL92 in January 2002, noting that DC had even more functionally illiterate (Level 1) adults than Mississippi, though Mississippi took the prize in Level 2 ("basic") reading capability. To a first approximation, functional illiteracy corresponds to the more widely used "Below Basic" category of the NAEP scoring system, which is used in many different analyses of DC's critical education problems. These difficulties are common among black (and Hispanic) kids raised in urban environments across the US, and generally traceable to the lack of education of their most often single parents. We also revisited this scoring issue in depth in our critique of the quantitative goals set in the recent DCPS "Strategic Plan" and its failure to compare DC's poor showing to that of other mostly black, urban schools. "Below Basic" scores between 40 and 50% are not unusual in these systems. We are further convinced that these horrendous scores get worse as the share of black kids in a school increases, and counter-cultural influences become the norm. Sampling that fails to consider the effects of minority-concentration in urban schools may well underestimate illiteracy among their graduates.
Is it likely that 25% of the DC workforce is functionally illiterate?
As noted earlier, the Kairos Technical Appendix purports to provide a definitive distribution of this "below basic" cohort of struggling adults in DC, assuming that the national proportions apply as well to the national capital, whose "inner city" public schools have long been roughly 90% black. In fact, we conclude they underestimate illiteracy in the workforce for three reasons (which readers may well consider quibbling):
o first, if one digs back into the 1997 Reder Study (available on the net), he quite properly provides "scatter plots" of the NAAL92 data, comparing those "observed at Level 1" (i.e., the test data) with his own "predicted at Level 1" (from his synthetic estimates derived from his regression analysis). While the sympathetic analyst will be favorably impressed by the correlation, NARPAC skeptics find two concerns, which we try to illustrate in the graphic below. To begin with, there is (not surprisingly) significant scatter (white band), such that if the observed (tested) percentage was 37%, the Reder predictions ranged from roughly 22% to 44% (i.e., uncentered). But of possibly more interest, or in keeping with NARPAC's bias, the "linear regression line" through the data (green line) is not coincident with a strictly linear (and equal) relation between predicted and observed (red line). This departure with increasing illiteracy is consistent with NARPAC's "counter-cultural factor" noted above. We suspect that when Reder's "predicted" value gets up to 37% of DC's 469,000 adults, the "observed" will have risen to perhaps 42%. This increases the predicted victim pool from 170,700 to 198,800.
o second, the consultants claim that 33% of Level 1 adults are over 65 and thus beyond the workforce. Using the adults with a high school degree or less as the metric, only 21% of this slice is retired nationally, and 23% in DC (using Census 2000). We would thus raise DC's (readjusted) workforce-aged share from 114,000 to 153,100.
o third, Kairos sets the "in the laborforce" fraction of Level 1 adults at 38%, or 43,000. While this appears to have been the valid number on the NAAL92 survey, the equivalent figure for NAAL03 is reported by the NCES "first look" to have risen to 45%. Applying this higher percentage to NARPAC's higher number of working-age illiterates brings the total number of employed functional illiterates to 68,900: a 60% increase (from 43,000), and for the worse!
Is such a complex analytical process required to generate such an estimate?
There may be some value to overwhelming the casual observer (and holder of the purse strings) with detailed and generally unintelligible analysis. Compared to the "Rafuse Report" conjured up some years ago to help justify a phony "structural imbalance" in DC's local finances for its CFO , the Reder effort is relatively harmless. The problem is that it includes several strictly "second-order terms" (at least for DC's single-issue case) such as: disabilities, language, immigrants, part time workers, occupations, and workforce participation (most of which are effects of, not causes of illiteracy). We have tried, perhaps unsuccessfully, to demonstrate this on the complex charts below, based on the Kairos 8-Ward analysis. On tyhe left, each "contributing factor" in estimating DC illiteracy is plotted separately (i.e., percent disabled, etc.) versus the total illiteracy score (i.e., % adults below Level 1) for each of the eight wards. The shallow sloping lines contribute little to the answer, while the heavier, steeper lines (red and blue) are the "first order terms". The trends in those two primary terms are graphed separately to the right
The lefthand chart also shows a) the linear correlation (thin green line) between 'Level 1' and b) 'Level 2' reading capability (which appears pre-ordained by the methodology), and a seemingly inevitable correlation between reading ability and enjoying a professional occupation. Neither is particularly relevant to the point being made.
On the other hand, we reiterate that this methodology ignores illiteracy's primary cause: lack of parents and/or parental education. In their place are substituted two largely irrelevant, albeit im- plicit, postulates:
o illiteracy is caused by having an inadequate education (oh, wow, imagine that!)
o illiteracy is caused by being black or brown (there is no statistical proof whatsoever that missing or seriously undereducated white or yellow parents produce smarter offspring than missing or seriously undereducated brown and black parents).
Is DC Gaining on the Problem?
Another question worth asking is whether or not illiteracy is on the way out as less well educated senior citizens die off and young school/college graduates enter DC's work force. To a first approximation, this should be indicated by the shifting fraction of poorly educated adults by age group. To NARPAC's surprise, however, this indicator appears to be quite different between a high-minority urban cluster like DC, and for the US as a whole.
One of the most startling abnormalities in comparisons between the nation as a whole and our capital city is the widely different apparent value of a high school degree: a topic not addressed in any of the relevant reports referenced here. This is illustrated on the right hand chart below which shows each "layer" of educational achievement for the adult population of each the US (left tall bar) and DC (right tall bar). These are compared to the fraction of each population currently judged to be functionally illiterate (outer red bars).While the DC distribution is significantly skewed towards better educated adults (bigger layers of college education), the functionally illiterate bar reaches virtually as high as total number of adults with high school diplomas (dark green). For the US as a whole, the illiterate share reaches only slightly beyond the (light green) band of those not finishing the twelfth grade (G-12-). To a first approximation, anyone in DC with a high school degree (or equivalent) is probably functionally illiterate. In the rest of the US as a whole, only those who failed to complete high school are likely to be illiterate. This is an extraordinary condemnation of urban public school systems.
The chart pair on the left side of the cluster above indicates that there seems to be little reduction US-wide (upper set) in the proportion of the adult population saddled with less than a school diploma. It is stable at about 13% in each age group and suggests little change is forthcoming at the national level. For DC, (lower set) the outlook for reducing illiteracy is somewhat more encouraging. Here, literacy depends on the number of kids going beyond high school into college, and there are minor improvements in each younger age group. Hence while the US trend depends on reducing the share of high school dropouts, the DC trend will only be encouraging if the share of the population going beyond high school continues to improve. This apparent condemnation of urban public school systems is not highlighted in the DCSEA effort.
Does illiteracy keep adults out of the workforce?
According to the DCSEA's call to arms against illiteracy: "employment and earnings are directly tied to an individual's educational achievement". While this statement appears, intuitively, to be above reproach, it is remarkably difficult to demonstrate in overall statistics. The cluster of simple charts below indicates this dilemma. Six sets of comparisons are made between DC (with its nearly 37% adult illiteracy rate) and the US as a whole, with its significantly lower 21% functional illiteracy:
o While DC has a slightly larger fraction of adults who did not finish high school, it has a substantial advantage among those who went through college and beyond:
o There are actually a larger portion of DC's adults in the labor force than across the US, though DC's unemployment rate is, to be sure, somewhat higher;
o Moreover, there are relatively more workers above the poverty line than US-wide, and a smaller share of unemployed that are below the poverty line;
o DC's worker mobility is perhaps 1% below the nation as a whole (27% work outside their home "county" rather than 28%;
o There are tiny percents fewer DC residents with disabilities both employed and unemployed;
o But only 36% of DC's next crop of adults are being raised by two parents, while across the US, the norm is 68%. And this factor, by itself, may explain why the value of a high school diploma in DC (and other cities) is far lower than across the nation as a whole.
Does illiteracy express itself as a racial issue?
Yes, it does, but it is far too easy to confuse cause and effect. Skin color does not effect reading comprehension, but it does tend to identify patterns in family relations and the culture passed on from parents to their kids. As NARPAC has insisted elsewhere , "Schools don't make dumb kids, dumb parents do!". But the chart set below does show how DC's adult residents (all vertical scales) cluster at the two ends of the earning and education scale. Household income distribution (upper left) puts the black and Hispanic "norm" in the $25-$50K area, while the cumulative income for the city is about 40% white or Asian (upper right) with a relatively small share of the population.
While the proportions of each age group are roughly the same for each of the seven educational attainment categories (lower left), it is abundantly clear than blacks and Hispanics dominate the lower education levels while whites and Asians are a clear majority of college attendees (lower right). Not emphasized here is the significant difference in both distributions between DC and the US average. A larger share of DC's adults without college education are high school drop-outs, (44% v 35%) but DC has a very significantly higher share of residents with college degrees (45% vs 25%). Similarly, DC has over four times as many blacks, but a lower share of Hispanics. Substantial departure from the 'norms' can produce greater 'scatter' in estimating results.
Where do DC's functional illiterate live?
This question is answered in spades by the DCSEA consultants' report meticulously re- constructed by Ward, and summarized in the chart below, which arranges the eight wards in order of declining literacy The green bars indicate the number of adults (16 and over) in each ward, while the blue bars indicate the number of registered voters (18 and over). The yellow bars indicate the number of working adults, and finally, the red bars show the number of adults that are not estimated to be functionally illiterate. In short, there are fewer literate adults than workers in six of the eight wards, and possibly more interesting, as many as 52% of DC's regular voters in those Wards are functionally illiterate!
What are these illiterate DC adults doing in the workforce?
The DCSEA report makes it clear where it thinks DC's working and non-working illiterate live, but one can only speculate as to what the "working illiterate" are doing. Available Census statistics provide information on the working skills of all workers by where they reside, and other statistics on how many working residents work outside their home "county" (or state). But what skills go outside DC to work are not provided. As a gross case in point, NARPAC concluded earlier in this analysis that DC could well harbor as many as 68,900 working illiterate. But Census also estimates that 66,900 DC workers are employed outside the District boundaries. Is it possible that these less skilled workers are all working at menial jobs in the suburbs while living in the city? Possible, but surely not probable, since most low wage earners are forced to live well outside the city limits despite the best intentions of DC's affordable housing advocates. Furthermore, one of the most heavily "out-trafficked" bridges from DC leads wealthy DC residents to high-paying day jobs in the Virginia suburbs. NARPAC takes a wild guess that no more than 25% of DC's illiterate workers are employed outside District limits.
But regardless of where they work, what they do is probably easier to discern: they are almost certainly in the lower paying jobs. The chart to the left here shows the distribution of DC job skills by gender, and between black residents and all the rest. Some 32,000 workers earn less than $25,000 and another 94,000 workers earn less than $40,000. In these two categories, 48% of the workers are females, and 61% are black, typical of the biases in lower income occupations. On the other hand, some 32,500 of these jobs are in office administration and sales, of which a good portion are certainly government workers. Does this confirm some of your darkest suspicions?.
The bulk of the DCSEA State of Adult Literacy report provides a lengthy rehearsal of all aspects of why and how DC's literacy rates should be improved. Here are the generally self-explanatory, and embarrassingly platitudinous titles and subtitles of that report. Its authors may have achieved a "basic" level of reading, but surely a "below basic" capacity for concise report writing. Note the emphasis is entirely on remediating the current pool of illiterates rather than on lowering the creation of new ones :
For general discussion of issues concerning public education in DC, and DCPS facilitiesplease refer back to the previous chapters.
For more detailed information on relevant reports on DC public education and related topics please refer ahead to the next chapter.
Did you find this of interest?
© copyright 2007 NARPAC, Inc. All rights reserved