Skip to content

STAR Test Scores by Grade: Scaled Score and Percentile Guide

Last Updated on 18 November 2026 by Editorial Team

Last updated: 17 November 2026 · Applies to the 2026–2027 school year · Reviewed against Renaissance score documentation and state education department guidance

The short answer. A STAR report gives you several numbers and only two of them matter for most purposes. The scaled score is the measurement, and it rises across a child’s whole school career. The percentile rank places that measurement against other children in the same grade nationally. Everything else on the report — grade equivalent, NCE, instructional reading level, benchmark category — is derived from those. A percentile rank of 40 is the threshold most schools use to mean “on track”.

What the STAR tests are

STAR is a family of computer-adaptive assessments published by Renaissance Learning. Three members of the family account for almost all school use:

  • Star Early Literacy — pre-reading and early reading skills, typically pre-kindergarten through the early part of grade 3.
  • Star Reading — reading comprehension and vocabulary, typically from the point a child can read independently through grade 12.
  • Star Mathematics — mathematics across the same span.

They are screening and progress-monitoring instruments, not end-of-year accountability tests. Schools usually administer them two to four times a year, often autumn, winter and spring, and use the results to decide who needs intervention, whether that intervention is working, and how children are growing. A single STAR score is a snapshot taken for an operational purpose. It is not a verdict.

Each test takes most children about 20 minutes, and because it is adaptive, no two children see the same set of questions.

Why the test gets harder, and why that is good news

STAR is computer-adaptive. The software chooses each question based on how the child answered the previous ones: a correct answer leads to something harder, an incorrect answer to something easier. The test converges on the level where the child is getting roughly half the questions right.

Two consequences follow, and both surprise parents.

First, a child who finds the test hard is usually doing well. The algorithm is deliberately pushing them to their limit. A child who reports that the questions were easy may have been on an easier path. Difficulty experienced is not a signal of performance.

Second, the number of questions answered correctly is meaningless. There is no “score out of”. A child who answers 50% correctly at a very high difficulty level will score far above a child who answers 90% correctly at an easy level. If your child comes home and says they got lots wrong, that is the test working as designed.

The two score scales, and which one you are looking at

This is the most confusing part of a STAR report, because Renaissance has been moving between two scales and both are still in circulation.

The older Enterprise scale runs from 0 to 1400 for Star Reading and Star Mathematics, and from 300 to 900 for Star Early Literacy. The newer Unified scale was created so that a single scale applies across all three assessments, letting a child move from Star Early Literacy to Star Reading without a break in the measurement. On the Unified scale, the overall span is 200 to 1400: Star Early Literacy occupies 200 to 1100, and Star Reading occupies 600 to 1400.

Assessment Enterprise scale range Unified scale range
Star Early Literacy 300–900 200–1100
Star Reading 0–1400 600–1400
Star Mathematics 0–1400 Shares the 200–1400 span

Why this matters to you. The same child, on the same day, has a different-looking scaled score on each scale. A score of 1075 means nothing until you know which scale it is on. Reports from different years, or from a district that has switched scales, are not directly comparable. Before you compare two STAR scores, check that both are on the same scale — the report usually labels it, and if it does not, ask the school. Where a report shows a Unified scale score, the percentile rank is derived from that Unified score.

Every score on the report, and what each one is for

Score What it is Use it for
Scaled score (SS) The underlying measurement, calculated from the difficulty of the questions answered and the pattern of responses. The same range applies to all children, so it can be compared across grade levels. Tracking growth over time. This is the number to watch across a school career.
Percentile rank (PR) The percentage of children nationally, in the same grade, who scored the same as or lower than your child. Runs from 1 to 99. Answering “where does my child stand?” This is the number for comparison.
Percentile rank range The band within which the child’s percentile rank would probably fall if they retested shortly afterwards. Understanding how precise the percentile rank is. Most parents never look at this and should.
Normal curve equivalent (NCE) A norm-referenced score similar to percentile rank but on an equal-interval scale, so differences between any two points mean the same thing. Arithmetic. Averaging or subtracting percentile ranks is statistically invalid; NCEs can be averaged.
Student growth percentile (SGP) How the child’s growth compares with that of children who started from a similar score. Runs from 1 to 99. Judging progress rather than position. A low-scoring child can have a very high SGP.
Grade equivalent (GE) The grade and month at which the median child achieves this scaled score. Very little. See the warning below.
Instructional reading level (IRL) A criterion-referenced score: the highest reading level at which the child comprehends at least 80% of the material with assistance. Choosing books and instructional materials.
ATOS 2000 The scaled score expressed on Renaissance’s own readability scale, similar but not identical to the Lexile scale. Matching the child to books by readability.
Benchmark category A band such as “At/Above Benchmark” or “On Watch”, assigned by comparing the score with a cut score. Knowing whether the school considers the child on track. Note the cut scores are set locally.

Percentile rank, stated precisely

A percentile rank of 85 means your child performed as well as or better than 85% of other children in the same grade nationally. Renaissance’s own documentation describes the percentile rank as the best measure of a child’s level of achievement relative to other children in the same grade.

Three properties of percentile ranks cause most of the confusion around them:

  1. A percentile rank is not a percentage correct. They are entirely different quantities that happen to share a scale from 1 to 100. On an adaptive test the percentage correct is approximately 50% for everybody by design.
  2. Percentile ranks are not evenly spaced. Children bunch up in the middle of the distribution, so a small change in scaled score near the median moves the percentile rank a long way, while the same change near the extremes barely moves it. A child who gains 30 scaled score points might move 12 percentile points if they started at the 50th, and 2 points if they started at the 95th.
  3. Percentile ranks are grade-referenced here, not age-referenced. Unlike ability tests such as the CogAT or NNAT, which compare a child against others of the same age, STAR compares against others in the same grade. A child who is young for their grade is competing against older classmates.

Because of point 2, percentile ranks must not be averaged or subtracted. If a school tells you a child “moved up 10 percentile points”, that is a meaningful direction but not a meaningful quantity. The NCE exists precisely for this purpose.

Percentile rank to NCE conversion chart

Normal curve equivalents are built so that the distance between any two scores means the same thing anywhere on the scale. They are anchored so that NCE 50 equals the 50th percentile, and the scale is stretched by a factor of 21.06 standard-deviation units. The conversions below are calculated from that definition.

Percentile rank NCE Percentile rank NCE
99 99.0 50 50.0
95 84.6 40 44.7
90 77.0 30 39.0
85 71.8 25 35.8
80 67.7 20 32.3
75 64.2 15 28.2
70 61.0 10 23.0
60 55.3 5 15.4
50 50.0 1 1.0

Look at the top and bottom rows to see the point of the exercise. Moving from the 90th to the 95th percentile is 5 percentile points but 7.6 NCE points. Moving from the 45th to the 50th is also 5 percentile points but only about 2.6 NCE points. The second move is a much smaller real gain, and only the NCE shows that.

The percentile rank range, and why it should change how you read the report

Renaissance publishes a percentile rank range alongside the percentile rank, and it is the most underused number on the report. It indicates the statistical variability in the percentile rank: a child with a percentile rank range of 32 to 59 would be likely to score within that range if they took the test again within a short period, meaning roughly four to six weeks.

Read that example again, because it is Renaissance’s own illustration and it is startling. A single reported percentile rank can carry a plausible range of nearly 30 points. That is not a flaw in the test; it is the normal precision of a 20-minute adaptive screener, and every brief assessment has something similar.

What to do with this. When the school reports a percentile rank of 45, read it as “somewhere in the middle, probably”. When a child moves from the 45th percentile in autumn to the 52nd in winter, that is almost certainly not a real change — both numbers sit inside each other’s ranges. Real signals look like several administrations trending the same way, or a move of 20 percentile points or more. Ask for the percentile rank range, and ask for the scaled score history rather than a single result.

Benchmark categories and the percentile rank 40 convention

Most schools convert the scaled score into a small number of categories, with names such as “At/Above Benchmark”, “On Watch”, “Intervention” and “Urgent Intervention”. These categories come from cut scores, and the cut scores are set by the state, the district or the school — not by Renaissance as a national standard.

The most widespread convention sets the benchmark at a percentile rank of 40. Several state education departments publish it in exactly those terms: children scoring above a percentile rank of 40 will likely meet end-of-year performance goals. That is the origin of the “40th percentile” threshold you will hear in parent meetings.

Two qualifications matter. First, a percentile rank of 40 is below the national median, so “at benchmark” does not mean “above average” — it means “likely to reach the end-of-year goal”. Second, because the threshold is expressed as a percentile but applied as a scaled score, the scaled score that corresponds to it changes with grade and with the time of year.

Example benchmark cut scores

The table below shows published cut scores at a percentile rank of 40 from one state education department, for the year stated. It is included to demonstrate the shape of these thresholds — how they rise across grades and across the three testing windows — not as a national standard.

Assessment Grade Autumn screening Winter screening Spring screening
Star Early Literacy K 690 730 769
Star Early Literacy 1 752 790 828
Star Early Literacy 2 868 893 918
Star Early Literacy 3 939 955 970

Source: Nebraska Department of Education, Star Suite cut scores, 2022–23. Threshold set at percentile rank 40.

Do not use this table for your child. These figures are specific to one state, one year and one scale version. Cut scores are revised, scales have changed, and your district may set a different threshold entirely. The correct source for your child’s benchmark is the cut score table your own state or district publishes for the current year, and the school is required to be able to tell you which one it used. We publish this example because the structure is genuinely useful to understand, and because the alternative — inventing a national table — would be worse than useless.

Notice what the pattern shows: the autumn threshold for grade 1 is above the spring threshold for kindergarten, and the gap between windows narrows as children get older. That narrowing is real and important. Early reading growth is extremely rapid, so a child a few months behind in kindergarten is a long way behind in scaled score terms, while the same gap in grade 3 represents less distance. It is the main statistical reason early intervention is emphasised so heavily.

Student growth percentile: the number that answers “is my child improving?”

A percentile rank tells you where a child stands. It does not tell you whether they are making progress, because a child can improve substantially and stay at the same percentile if their peers improved too.

The student growth percentile solves this. It compares a child’s growth with the growth of children who started from a similar score, and runs from 1 to 99 with the same style of interpretation as a percentile rank: lower numbers mean less relative growth, higher numbers mean more.

The crucial property is that SGP and percentile rank are independent. All four combinations occur and each means something different:

Percentile rank SGP What it means
Low High Behind, but closing the gap faster than similar children. The intervention is working. Keep going.
Low Low Behind and not catching up. This is the combination that warrants a conversation about changing the approach.
High High Ahead and extending the lead.
High Low Ahead but growing slowly. Often means the child is not being stretched, and is the pattern most frequently missed, because the headline percentile looks reassuring.

If you ask one question at a parents’ evening about a STAR report, ask for the SGP alongside the percentile rank.

Grade equivalent, and why you should mostly ignore it

The grade equivalent is the score most likely to be misunderstood, and the one that causes the most unnecessary alarm and unnecessary complacency.

A grade equivalent of 5.2 does not mean your grade 3 child can do grade 5 work. It means your child’s scaled score is the score a median child in the second month of grade 5 would obtain on this test. Those are very different claims. The grade 3 child reached that score by answering grade-3-appropriate material unusually well; they have not been assessed on grade 5 content, and extrapolating from the grade equivalent to curriculum readiness is not supported.

The same applies downward. A grade equivalent of 2.1 for a grade 4 child does not mean they read like a typical grade 2 child; it means their score matches that benchmark on this instrument.

Grade equivalents also become less meaningful the further they sit from the child’s actual grade, because they are extrapolations. For instructional decisions, the instructional reading level is the appropriate criterion-referenced score, and for comparison the percentile rank is the appropriate norm-referenced one.

Five common misreadings

  1. Treating the scaled score as a percentage. A Star Reading scaled score of 450 is not 45%. On the Enterprise scale it is a point on a 0 to 1400 continuum.
  2. Comparing scaled scores across scale versions. An Enterprise score and a Unified score are different measurements. Check the label first.
  3. Reading “at benchmark” as “doing well”. The common threshold sits at the 40th percentile, which is below the median.
  4. Reacting to a single test. With a percentile rank range that can span nearly 30 points, a one-off movement is usually noise. Ask for the trend.
  5. Using the grade equivalent to decide what the child should be taught. Use the instructional reading level for that.

Frequently asked questions

What is a good STAR score?

There is no single answer, because the scaled score has to be read against grade and time of year. In percentile terms: above 40 is “on track” by the most common convention, above 50 is above the national median for the grade, and above 75 is comfortably strong. Below 25 generally triggers intervention in most schools. If you want one number to watch over time, watch the scaled score; if you want one number for context, use the percentile rank.

My child’s scaled score went up but the percentile rank went down. How?

Because percentile rank is relative. Your child improved, and the comparison group improved more — or your child moved into a new grade’s norms, where the bar is higher. This pattern is common between a spring test and the following autumn. It is not a contradiction and usually not a cause for concern; check the SGP to see whether growth was below par.

Can my child practise for STAR?

Not in any useful sense, and schools do not want them to. It is adaptive, so there is no fixed content to prepare, and it is designed as a measurement of what the child can already do so that teaching can be targeted. Practising to inflate a screening score would make the instrument worse at its only job. The one legitimate preparation is making sure the child knows how the interface works and understands that difficult questions are expected.

Why does my child take it three times a year?

Because the purpose is monitoring, not ranking. Three data points let the school fit a growth trajectory, check whether interventions are working, and catch a child who is slipping before the end of the year. One annual score could not do that.

How long is the test and can my child take a break?

Most children finish in about 20 minutes. Policies on pausing vary by school and by assessment; ask your school. If your child has an accommodation on an IEP or similar plan, it generally applies to STAR as it does to other assessments, and you should confirm with the school that it was applied.

Is STAR the same as the state test?

No. It is a commercial screening and progress-monitoring tool. In some states, however, STAR scores are linked to the state scale so that families can see an equivalent — Florida, for example, publishes equivalent scores linking the Star Unified scale to the state’s own scale. An equivalent score is a translation, not a prediction, and the linkage is approximate at the extremes of the scale.

Where our numbers come from

The NCE conversions on this page are calculated from the standard definition of the normal curve equivalent scale. All scale ranges, score definitions and threshold conventions are taken from Renaissance’s published score documentation and from state education department guidance, listed below. These are technical and policy documents rather than journal articles and do not carry DOIs; the publishing body and year are given so each can be located.

  1. Renaissance Learning, Inc. Star Reading score definitions [technical documentation]. Wisconsin Rapids, WI: Renaissance Learning. Source for the Enterprise scaled score range of 0 to 1400, the definition of percentile rank, and the definition and worked example of the percentile rank range (a child with a range of 32–59 being likely to score within that range on retesting within four to six weeks).
  2. Renaissance Learning, Inc. (2024). Understanding FAST: Parent quick guide. Wisconsin Rapids, WI: Renaissance Learning. Source for the Star Unified Scale description, the statement that a single percentile rank is reported based on the Unified Scale score, and the linking of the Unified scale to a state scale.
  3. Nebraska Department of Education. (2023). Star Suite cut scores [state approved assessment documentation]. Lincoln, NE. Source for the Unified scale span of 200 to 1400, the Star Early Literacy range of 200 to 1100, the Star Reading range of 600 to 1400, and the example cut scores at percentile rank 40 reproduced above.
  4. Nebraska Department of Education. (2024). Star Suite Reading request for information, 2024–2025. Lincoln, NE. Source for the Unified score scale rationale and the percentile rank definition.
  5. Wisconsin Department of Public Instruction. Star assessment data in WISEdash: Scoring [state guidance]. Madison, WI. Source for the Star Reading and Star Mathematics scaled score range of 0 to 1400, the note that all other Star scores are derived from the scaled score, and the definition and 1–99 range of the student growth percentile.
  6. Lakeland School Corporation. Star Reading score definitions [district handout]. Source for the normal curve equivalent definition as an equal-interval norm-referenced score, and for the ATOS 2000 scale description.
  7. Renaissance Learning, Inc. Star Early Literacy, Star Reading and Star Mathematics [score and report documentation]. Source for the instructional reading level definition as the highest level at which a student comprehends at least 80% of material with assistance.

A note on what we do not publish. Several sites publish “STAR score charts by grade” giving a specific scaled score for each grade and percentile. Those tables are generated from Renaissance’s proprietary norms, which are revised periodically and differ between the Enterprise and Unified scales, and a chart that does not state its scale version and norm year is not usable. Rather than reproduce numbers we cannot verify, we explain the scales, give the conversions that are mathematically derivable, and show one clearly sourced example of a real cut score table. If your school has given you a figure that does not match something here, tell us at the address below and we will check it.

Found an error on this page? If a number here does not match what your school has given you, or you think something is wrong, write to the editor at Editorial@schooltest.org. We answer correction requests and we date every page so you can see when it was last checked.