What to Look for in a School Improvement Plan

A school improvement plan is the public document in which a school identified for improvement explains what is going wrong, what it will change, and how it will know whether the change worked. The most useful way to read one is not to check whether every required section is present, but to ask whether the evidence in the plan actually justifies the actions it proposes. State review rubrics — the scoring guides agencies use to grade these plans — make that standard explicit.

This article explains what reviewers look for, drawing on two published rubrics as evidence, and offers a checklist that parents and community members can apply at a school or district meeting. It also flags which New York State details still require verification against current New York State Education Department (NYSED) guidance.

What a school improvement plan is

Federal education law requires states to identify schools where student groups are persistently underperforming and to require those schools to plan for improvement. The two categories most readers will encounter are Comprehensive Support and Improvement (CSI), generally applied to the lowest-performing schools and to high schools with very low graduation rates, and Targeted Support and Improvement (TSI), applied to schools where specific student groups lag behind their peers. A school improvement plan is the document produced in response.

Plans are reviewed — by the district, the state, or both — and the review typically follows a published rubric that rates each section. That is why the wording of a plan matters: reviewers are asked to judge the quality of the reasoning, not merely the presence of a heading.

How reviewers separate strong plans from weak ones

The clearest published example available here is the school improvement plan rubric from the District of Columbia’s Office of the State Superintendent of Education (OSSE). It uses a four-point scale, from Weak (1) to Strong (4), applied section by section, beginning with the needs assessment.

On the needs assessment, the rubric asks whether a needs assessment was conducted to determine the nature and cause of underperformance, whether a narrative of the results was included, whether that narrative outlines the problem areas that the plan will address, and whether multiple data sources were used. A plan lands at the weak end when challenges are mentioned without any data, when some data appear but not enough to draw conclusions about school performance, or when no needs assessment is mentioned at all. It reaches the strong end when the narrative summarizes the assessment results, identifies areas of challenge that must be addressed, and rests on multiple data sources.

On goals, the rubric’s weak end describes goals that are not S.M.A.R.T. and that are not linked to specific subjects, non-academic areas, or other defined targets. Its strong end describes goals and objectives that clearly address the school challenges identified in the needs assessment. In other words, the goal section is graded on its connection to the diagnosis, not on how ambitious it sounds.

One caveat about that document: the OSSE rubric was written to support reviews under the District’s flexibility waiver to the Elementary and Secondary Education Act (ESEA), a pre-2015 mechanism that the Every Student Succeeds Act later replaced. It is best read as an illustration of how reviewers reason, not as a statement of current District of Columbia policy.

A second example, the School Improvement Plan Quality Review rubric published by Pennsylvania’s Intermediate Unit 13 (IU13), is useful mainly for what it says about process. According to that document, school committees are meant to read the planning guidelines and the rubric before drafting each section and again afterward to find revisions. District leaders are meant to use it during the “Ready, Set, and Go” phases of plan development, give feedback on each section, and approve only plans that meet the expected criteria before those plans go to the board and to the Pennsylvania Department of Education. Facilitators use the same rubric as a coaching tool. The rubric is aligned to Pennsylvania’s School Improvement Plan Guidelines and the Future Ready Comprehensive Planning Portal.

Neither rubric is New York’s instrument, and the two come from different states and different policy eras. What they share is a logic that carries across states: reviewers look for a documented diagnosis, priorities that follow from that diagnosis, and goals that can be checked later.

What to look for: a component checklist

The following items are framed as questions a reader can ask. The first three follow directly from the criteria in the rubrics described above; the rest reflect standard expectations for how a plan is written and used.

1. A needs assessment built on more than one data source

Look for a short narrative, not just a table. It should name the specific problem areas — which grades, which subjects, which student groups — and it should draw on several kinds of evidence, such as state test results, graduation or attendance data, course-taking patterns, discipline records, and survey or climate data. A plan that cites a single test score is thin evidence for a multi-year strategy.

2. Priorities that follow from the data

If the needs assessment identifies chronic absenteeism as the main barrier but the plan’s priorities are all about instructional materials, something has broken down between diagnosis and treatment. Reviewers are explicitly asked whether the narrative connects the problem areas to the priorities the plan will pursue.

3. Goals that are specific and measurable

S.M.A.R.T. goals — specific, measurable, achievable, relevant, and time-bound — name a population, a measure, a baseline, and a target date. “Improve student achievement” is not a goal that can be evaluated. “Increase the share of ninth graders passing Algebra I from 62 percent to 75 percent by June” can be.

4. Strategies with owners and timelines

Each strategy should say who is responsible, what changes in practice, when it starts, and what resources it needs. A list of activities with no owner is a wish list.

5. A monitoring plan

Ask how and how often the school will check whether the strategy is working, and what it will do if the data do not move. Plans that only describe the first year of activity are harder to correct mid-course.

6. Families and staff as participants

Look for evidence that teachers, families, and students had a role in shaping the plan — meeting dates, survey results, or a description of who served on the planning team — rather than a note that the plan was shared after it was finished.

7. Plain-language publication

A plan that families cannot find or read cannot be used. Check whether it is posted on the district or school website, whether the version posted matches the version submitted, and whether key sections are available in languages spoken in the community.

Reading the numbers in a plan without being misled

Plans are full of percentages, and percentages hide their denominators. A claim that proficiency rose from 20 percent to 40 percent means something very different in a group of 10 students than in a group of 400. Before treating a number as a trend, identify the entity (which school, which grade, which student group), the period or cohort, the metric definition, and the population it covers.

Three specific cautions are worth carrying into any plan review:

• Small groups are unstable. Year-to-year swings in a subgroup of a few dozen students can reflect who happened to be in the room rather than a change in instruction. Many reporting systems suppress small-group results for exactly this reason; a blank or missing cell is not the same as a zero.
• Definitions change. Test formats, proficiency cut scores, graduation cohort rules, and accountability formulas are revised over time. A comparison across years is only meaningful if the metric definition held steady, or if the plan explains the change.
• Association is not cause. A plan may note that reading scores fell while absenteeism rose. That pairing is a reason to investigate, not proof that one caused the other.

None of this makes the data unusable. It means the plan should say what it is comparing and acknowledge what the numbers cannot establish.

Questions families can ask

• What exactly was the school identified for, and which student groups or indicators drove the identification?
• Which data sources went into the needs assessment, and can we see the results for each grade and student group?
• What are the two or three priorities, and why were they chosen over the alternatives?
• What is the baseline for each goal, and what target has the school set?
• Who is responsible for each strategy, and what resources have been committed?
• How will progress be reported during the year, and to whom?
• What happens if the data do not improve — who reviews the plan again, and on what schedule?
• How can families and students contribute to the next revision?

New York context: what this article can and cannot confirm

New York operates its own accountability and improvement system under federal law, with its own designation categories, planning templates, review timelines, and public reporting through the state’s school report cards. NYSED materials have referred to school-level and district-level comprehensive improvement planning documents, and to reviews conducted through the state’s school improvement process, but the exact current names, deadlines, templates, and scoring criteria for the 2026-27 school year are not established by the sources used for this article.

Verification gap: the two rubrics cited here are from the District of Columbia and Pennsylvania, not New York. Any statement about which New York schools must submit a plan, what the plan must contain, who approves it, or how it is scored should be confirmed directly against current NYSED guidance and the school’s own report card before it is relied on. Source-check date not recorded for New York-specific requirements; editorial verification required.

What does carry over is the review logic. Whatever the state’s template is called, a reader can still ask whether the diagnosis is documented, whether the priorities follow from it, whether the goals can be measured, and whether anyone has committed to checking the results.

Sources and methodology

This article is based on two published plan-review rubrics: the OSSE School Improvement Plan Rubric (District of Columbia) and the School Improvement Plan Quality Review rubric published by IU13 (Pennsylvania). Both were supplied as search results retrieved October 12, 2026, which is the source-check date for the rubric criteria described here. The OSSE rubric’s rating language is quoted and paraphrased from the needs assessment and goals sections of that document; the available excerpt of the goals section is truncated mid-criterion, so only the visible criteria are described. The IU13 rubric excerpt covers purpose and intended use rather than detailed scoring criteria.

General statements about federal identification categories, statistical caution, and plan-reading practices reflect established general knowledge rather than the two rubrics, and are presented as guidance rather than as findings from those documents. No New York school, district, or student data are reported in this article, and no illustrative numbers are attributed to any real school.