How Standardized Tests Are Graded: Percentiles Explained

Why Standardized Test Scores Are Not What They Seem

When a student receives a score of 1350 on the SAT or the 2nd grade on the Korean CSAT (수능), what does that number actually mean? Understanding how standardized tests are scored reveals not just the mechanics of grading but also important truths about fairness, measurement, and what these scores can and cannot tell us.

Csat Grade Kr

Raw Scores vs. Scaled Scores

Nearly every major standardized test converts raw scores to a scaled score. The raw score is simply the count of correct answers (minus any penalty for wrong answers, in tests that use this design). The scaled score is a transformed number that accounts for variation in test difficulty across administrations.

Why scaling is necessary: If the October SAT is harder than the March SAT, students who took the harder version would score lower despite equivalent ability. Scaling (also called "equating") adjusts for this, so that an 1100 in October represents the same ability level as an 1100 in March.

The SAT uses a scale of 400–1600, combining a 200–800 Evidence-Based Reading & Writing section with a 200–800 Math section. Raw-to-scaled conversions are test-specific and published in College Board score reports.

Percentiles: The Key to Understanding Your Rank

A percentile tells you what percentage of test-takers scored at or below your score. This is distinct from a percentage correct.

Percentile = (Number of test-takers with scores at or below yours / Total test-takers) × 100

Examples using 2024 SAT data: - Score 1500: approximately 96th percentile (scored as well as or better than 96% of test-takers) - Score 1200: approximately 74th percentile - Score 1000: approximately 40th percentile

Common confusion: a student scoring "in the 90th percentile" did NOT necessarily answer 90% of questions correctly. The percentile rank is purely about relative standing among test-takers.

Korea's CSAT (수능): Grade-Based System

The Korean College Scholastic Ability Test (수능, Suneung) does not report percentile scores directly. Instead, it uses a 9-grade system (등급) where each grade corresponds to a percentile band:

Grade	Cumulative Top %	Meaning
1	Top 4%	Exceptional
2	4–11%	Excellent
3	11–23%	Very Good
4	23–40%	Good
5	40–60%	Average
6	60–77%	Below Average
7	77–89%	—
8	89–96%	—
9	96–100%	—

The grade is determined by the student's standardized score (표준점수, pyojun-jeomsu) relative to the national distribution, not by a fixed cut-off. The standardized score formula is:

표준점수 = 20 × (raw score − mean) / standard deviation + 100

This means the average student always scores exactly 100 on the standardized scale, and the standard deviation is set to 20. A score of 140 means a student is 2 standard deviations above the national mean.

Percentile To Grade

The ACT: Composite Scores and Subject Scores

The ACT scores each of four sections (English, Math, Reading, Science) on a 1–36 scale. The Composite score is the average of these four, rounded to the nearest whole number.

ACT Composite	Approximate Percentile	US College Context
36	99th+	Perfect score; elite research university territory
33–35	97th–99th	Highly competitive selective universities
29–32	90th–96th	Many top 50 universities
24–28	74th–90th	State flagship universities
20–23	46th–74th	Most 4-year colleges
Below 20	Below 46th	Open-enrollment colleges

Curve Grading: How Grading on a Curve Works

"Grading on a curve" is frequently misunderstood. In the original sense, it means adjusting grades so the distribution matches a predetermined bell curve — most students receive C, fewer receive B or D, and very few receive A or F.

In practice, two common forms of curving: 1. Adding a fixed number of points to all scores (e.g., +10 points because the test was unusually hard) 2. Scaling to a target mean (e.g., ensuring the class average is 75%)

In standardized testing, the curve is replaced by equating — a more sophisticated statistical process that aligns scores across test forms.

The Limitations of Standardized Testing

Understanding how tests are graded also means understanding what scores cannot tell you:

• Coaching effects: SAT/ACT scores improve with intensive preparation, raising questions about whether scores measure innate ability or access to resources
• Score plateaus: The marginal difficulty of moving from 1450 to 1550 SAT is far greater than from 1000 to 1100, because the highest-scoring questions are designed for minimal discrimination at the top
• Subject vs. ability: The SAT Math section tests algebra and basic geometry, not calculus or statistical reasoning — a mismatch with STEM programs that assume far more advanced preparation
• International variation: A Grade 1 on the Korean CSAT reflects extraordinary competition in a highly uniform educational system; comparisons with ACT scores require careful contextualization

Practical Takeaways for Students and Parents

1. Focus on percentile rank rather than absolute score for college research — a school's middle-50% percentile range tells you more than a single target score
2. Subject-specific percentiles matter for specialized programs; a 760 Math SAT may be expected for engineering programs even if the overall score is in the 95th percentile
3. For Korean students, the grade system means small raw score differences at the grade boundaries can produce dramatically different outcomes
4. Score trends across multiple test attempts are as informative as the peak score — consistent improvement signals genuine learning

Csat Grade Kr