How "Improvement" Works in ESEA


Both houses of Congress have included requirements in the reauthorization of the Elementary and Secondary Education Act (ESEA), which includes Title I remedial education, to test all students annually in grades 3-8 in reading and math, even though only 15 states now have all that testing.

It also requires states to use the results of test scores, and scores alone, to "evaluate" and possibly sanction schools and districts which fail to make "adequate yearly progress" (AYP) toward the goal that all students reach the "proficient" level on state tests within 10 - 12 years. Additionally, schools will have to report separately for students who are low income, of a minority race, have a disability, or are of limited English proficiency, if there are enough such students in a school. If any one of those groups do not also make AYP, the school will be subject to sanctions. In each school, 95 percent of all students and 95 percent of studetns in each named group must be tested for a school to be eligible for AYP; in other words, if 95 percent are not tested, the school will face sanctions. A number of states now allow parents to opt their children out of testing, so this provision will effectively overturn those state laws.

In both House and Senate, if schools fail to show AYP in any one year, they will be identified for school improvement. At this point, if they do not show enough progress the next year, they are subject to "corrective action," and if they do not show enough progress the third year, they will be slated for "restructuring," which will require "at least one" of the following: reopen the school as public charter school, replace principal and all or most of staff, contract management to private company, or turn operation over to state. In addition, parents can take 40 percent of their child's
pro-rata share of Title I and use it for private tutoring from a state-approved tutoring agency.

The House and Senate differ on AYP in that the House will require all the groups within the school to make the same adequate yearly progress as the school is required to do and will make the decision based on year-to-year changes in test scores. According to a Senate staffer, the current Senate version would require at least a one-percent-per-year increase in scores for the named groups, and the increase would be based on a three-year average, not one-year changes.


1) Making a decision based solely on changes in test scores will result in many mistakes because test scores are estimates, not exact, and because many factors can cause rapid fluctuations in school level scores, such as having a few particularly high- or low-scoring students. This is particularly true of schools with fewer than 60 students at a grade level, which is a large percentage of the nation's schools. This situation is worse when looking at groups within schools because the numbers are smaller and the results even less accurate.

Some researchers have said that the results will be determined primarily by "luck." Others have pointed out that because of the statistical properties of test-score changes, districts will be more likely to have more schools showing "improvement" if the schools are segregated. It is almost impossible for any school, even any district, to make score gains every year. Thus, almost all schools in the nation will be labeled and start to face sanctions.

2) The emphasis on testing for accountability will increase pressure on schools and teachers to teach to the test, to reduce curriculum to test coaching, to eliminate courses not tested, to have teachers teach out of subject in tested areas, to focus on narrow remediation and drill: to do all the things that boost scores in the short term but fail to support real increased learning. The damage to curriculum and instruction will be most severe to schools serving low-income and minority-group students, but the accountability provisions are such that it will increase the problem in almost all schools.

Details of the Plan

Below is a detailed synopsis, with occasional comments in brackets [], written by FairTest's Monty Neill:

I worked from the House committee version first; I compared that with the House version sent on to the Senate, and found no differences in the relevant sections; I then read the Senate committee version and found some differences, which I have noted (and which no doubt make this a bit more confusing). Because the Senate is still considering the bill, I did not have the most recent information available on the Senate website. I thus talked with Senate staff, who informed me in particular about the Senate-House differences in "adequate yearly progress" (AYP).

I present the information in four parts: Testing, Adequate Yearly Progress or AYP, Improvement, Parental Options; and then offer more comments at the end.

[Finding the legislation: The bills are available on the web: S. 1 is at and H.R. 1 is at . In each case, look for current legislation search, enter bill number. You will see several versions of the bills. For the House, select the version forwarded to the Senate. There are several presentation formats, including Adobe and a text version that is directly under the opening paragraph of the main version of the bill and is easiest to read. Because the legislation is to amend existing law, it is presented in a format that looks sort of like: Sec. 111... amends Sec 1111 – the text is under the 4-digit number of the existing law, and I refer to the 4-digit version.]

I. Testing (appears under Section 1111, State Plans)

A. Continue with requirement (from 1994) that state must test once in each of grades 3-5; 6-9; 10-12, in reading and math, tests to be based on state standards. [Every state now tests, but many tests are not based on state standards.]

B. By 2004-05 (House) or 2005-06 (Senate) school year, introduce annual testing of all students in grades 3-8 in reading and math (in Senate, add science in 2007-08); and keep testing once in grades 10-12. [Only 15 states now test reading and math annually in all grades 3-8.]

C. Score and report by "levels": below-basic, basic, proficient, advanced; state decides the test and the meaning of each level (so "proficient" may vary greatly from state to state).

D. Disaggregate data by income, race, disability, LEP for adequate yearly progress purposes (House and Senate), add migrant and gender for general reporting purposes in House and AYP purposes in Senate.

E. Report the test results publicly, including disaggregation of data as in I.D. above.

F. Severe special needs students may be given alternate assessments, less severe may be given accommodations.

G. English language learners (ELL) who have been in U.S. 3 years or less do not have to take regular English versions of tests, but in House version [I did not see this in Senate] all ELL students will have to be tested by the LEA to determine proficiency in English starting in 2002-03 [I assume even in first year of being in US, as California now requires].

II. Adequate Yearly Progress (AYP)

A. Use test-results-ONLY as the basis for determining adequate yearly progress. House and Senate say also gather dropout data. [However, the actual AYP formula in each chamber does not include dropout data in the calculation, so there appears to be no penalty for pushing students out of school to make AYP test results look better.] Other information may be gathered and reported, but it may not change which schools are subject to sanctions.

B. House provides two and Senate one model for meeting AYP requirements, but the essence is that in 10-12 years from the base year (2002), ALL students must reach the "proficient" level on the state tests.
[I find the language unclear, but there <may> be an exemption for severe special needs students, also may be one for ELL students for their first three years in US schools, but I don't think so – that is, they too are expected to meet "proficiency" requirements, even though they may not have to be tested with the regular tests.]

C. 95 percent of students in a school and 95% of students in each identified group in a school (racial, etc.) must be tested for a school to be able to meet AYP. [My reading is that this does not exempt severe special needs or ELL, who presumably must be tested despite other language saying they do not have to be tested - or this is a glitch in the writing of the law. Like mandating more testing, this will override state opt out provisions where they exist.]

D. Essentially, every school, district, and state, and all the listed groups (racial, etc.) must make adequate annual progress toward the goal of all being proficient in 10 (Senate) or 12 (House) years. E.g., subtract percentage of students now at proficient or above from 100 percent of students, divide result by 10 (or 12), to get the minimum percentage gain required each year for all the categories.
The Senate refers to "measurable progress" and a timeline for the groups and allows the state to define how it will determine statistically whether AYP is met (unlike the House which provides two options).
The Senate also differs on progress for the named groups: While the House version requires all named groups to make the same adequate yearly progress as is required of the school as a whole, and if they do not, the school is subject to sanctions. The Senate version [according to a Senate staffer, this is not in the website version] instead would require named groups to progress at least one percentage point per year, based on three-year averages (i.e., 3 points in any three-year block).

E. At most (but not all) places the groups (racial, etc.) are mentioned, provisions state the data must be reported/used only if statistically reliable (tests in general are required to be reliable) and so long as the population is not so small that individuals could be identified. [Recent evidence suggests that most schools, never mind groups, would be exempt from having to meet AYP (and thus miss sanctions) if reliability is taken seriously – which it probably won't be unless states or districts fight].

III. "Assessment" and "Improvement" (Sec. 1116) – the schedule of assistance and sanctions in the House bill proceeds as follows (Senate schedule differs, see below), assuming legislation passes in 2001:

Year 1 (2002) - baseline data established using the tests mandated in the 1994 authorization (see I.A., above).

Year 2 (2003) - Identification for school improvement; House language reads: "Sec. 1116(b)(1)(A). Identification.–A local education agency [LEA] shall identify for school improvement any elementary or secondary school served under this part that– (i) fails, for any year, to make adequate yearly progress." [In other words, if a school does not improve from year 1 to year 2, they are marked as not making adequate yearly progress, and same will be true for year 2 to 3, 3 to 4, etc.] Assistance must be provided to such a school. For any such school, its students may transfer to another school in the LEA that has met AYP, if permitted under state law; LEA is to pay for transportation. If all schools in LEA need improvement, LEA is to try to work out cooperative arrangement with other LEAs. Schools will have to develop/revise improvement plans. The Senate has similar requirement (see notes below) but Senate also says that schools or LEAs which are not participating in Title I are not subject to the "school improvement" provisions of Sec.1116(c), school improvement, corrective action and reconstitution.

Year 3 (2004) - "Corrective action" begins for any school which did not meet AYP after the one year of assistance. The LEA must continue to provide assistance and the option to enroll in other schools, and "take at least 1 of the following corrective actions," which include replacing school staff, implementing new curriculum, decrease school management authority, appoint outside expert, extend school year or day, restructure internal organization of school. Senate has shorter list; also senate allows an additional year before corrective action starts, and thus before subsequent more severe sanctions kick in.

Year 4 (2005) - Restructuring: If after a year of corrective action, a school is not making adequately yearly progress, the transfer option continues and by the start of the next year (year 5 – 2006, in House) the LEA will implement 1 of the following (consistent with state law): reopen school as public charter school; replace principal and all or most of staff; contract management to private company; turn operation over to state. [House language is a bit confusing to me – the actual reconstitution I think takes place in year 5, but maybe in year 4 – language is a bit clearer in Senate as starting the next year. With the Senate delay of one year for corrective action, this would make the Senate's year for restructuring be Year 6, 2007. While I may not have the exact years, the sequence and meaning are clear.]
Duration: if any school identified for corrective action or restructuring makes AYP for 2 years, they are off sanctions.
State: ensures LEA's do what they are supposed to, intervene if needed; monitor test results to be sure all schools are making AYP.
State identifies LEAs - under a similar system, state can identify LEAs as needing improvement (if LEA for 2 continuous years don't make AYP), and if LEA's don't improve, state shall "take corrective action" including taking them over.

Notes on AYP and "improvement": It is highly likely that very many (maybe nearly all) schools will at some point fail to make AYP and so be labeled. Since two consecutive years' failure is needed before "corrective action," the fluctuations of test scores (and games people play, such as retention, push out) will make correction less likely – except that as schools get more students to "proficient," it will be harder and harder to make AYP.
Since states vary widely in what "proficient" means, those with higher "standards" are more at risk – making an incentive to lower the standards (unless state authorities want all their schools to get labeled, for example as an argument for vouchers....). It may be that the National Assessment of Educational Progress (NAEP) will be used to measure state progress (Senate requires it, in House it is one of 2 options), but I did not see anything to require states to align their definition of "proficient" with NAEP's (which would guarantee most all schools will soon "fail").

IV. Parental options [1116(d)]
For a school at the "restructuring" stage, the LEA will make supplemental services available to eligible students (all those in a whole school program, those getting targeted assistance in other schools) from a state-approved provider ("in accordance with reasonable criteria that it shall adopt") as selected by the parent(s). The LEA will contract with provider, including goals and timelines. Costs are not to exceed 40% of the school's per child allocation, unless that is insufficient, then funds from another part of ESEA also can be tapped. Note that in whole-school programs, currently the allocation is around $600 (varies district to district); I don't know about "targeted assistance" students but assume it is somewhat similar. The supplemental assistance shall continue until child complete highest grade in the restructuring school or until school not under restructuring.

Final comments

These policies are likely to affect many more schools than those which serve mostly poor kids, though those will be most heavily affected. Since all that "counts" is test scores (not even dropout data counts), pernicious practices will be reinforced. It will obviously greatly increase teaching to test and narrowing curriculum and instruction (so kids get nothing but drill to reading and math tests). It will foster grade retention and pushing students out of school. The error rates will be fantastically high and perhaps the grounds for litigation.

The Bush administration has touted vouchers, but it appears to have backed off. It is quite certain vouchers will not pass this Congress, although a trial voucher program will likely be approved for a limited number of districts (two per state). However, the accountability provisions, from the labeling and sanctions to the use of funds for private tutoring programs (including faith-based), lay the groundwork for a more intensive voucher effort in the future.

Currently (early June), consternation over AYP has erupted. The White House is worried particularly about the House version, fearing the consequences of having virtually all schools labeled, and a great many of them by 2003 and 2004 - before the next election. While this labeling may please voucher proponents, it carries more risk of angering suburbanites who like their schools. Thus, a push is on to alter AYP, at least to the Senate version which will name fewer schools and institute corrective action later (after the election). I would not be surprised so see further tinkering so that only schools which also have low overall scores – say the bottom 10 percent – are faced with corrective action. This would allow voucher proponents a line toward continuing to aggressively focus on urban districts. It also would enable states to focus "assistance" since virtually none will in fact be able to assist even 10 percent of schools, never mind the great majority.

Click here to find out how you can join the fight against the new testing plan.