{{ summaryHeading }}
{{ averageDisplay }}
{{ summaryLine }}
{{ scoreCountDisplay }} {{ passRateDisplay }} {{ medianDisplay }} {{ shapeBadgeText }} {{ outlierBadgeText }}
Scores Bands Center Outliers
Use one row per score; names stay hidden unless Show labels is enabled.
{{ scoreSourceHint }}
Enter the total points possible, e.g. 50 or 100.
points
One band per line: label, minimum percent, e.g. A,90 and B,80.
Band threshold preview:
{{ activeThresholdBand.label }}
Active band follows the current dominant score band, or the pass threshold before scores are available.
{{ band.label }} {{ band.range }} {{ band.note }}
Enter 0-100%; use the policy cut score, e.g. 50 or 60.
%
Use None for clean reports, Z-score for mean distance, or IQR for quartile fences.
Common review range is 2.0-3.0 standard deviations.
Use 5 or 10 for classroom views; accepted range is 1-50 percentage points.
percentage points
Use 0-3 decimals; 1 is usually enough for class summaries.
Leave off when copying or exporting a privacy-safe class summary.
{{ params.include_names ? 'On' : 'Off' }}
Affects exported diagnostics only; statistics and charts stay unchanged.
{{ params.keep_header_preview ? 'On' : 'Off' }}
Metric Value Note Copy
{{ row.label }} {{ row.value }} {{ row.note }}
Distribution bucket Count Share Cumulative Copy
{{ bucket.label }} {{ bucket.count }} {{ bucket.shareDisplay }} {{ bucket.cumulativeDisplay }}
Band Minimum Count Share Copy
{{ band.label }} {{ band.minimumDisplay }} {{ band.count }} {{ band.shareDisplay }}
Label Raw score Percent Band Review flag Copy
{{ row.label }} {{ formatRaw(row.rawScore) }} {{ formatPercent(row.percent) }} {{ row.bandLabel }} {{ row.flagLabel }}
Line Input Reason Copy
{{ row.lineNumber }} {{ row.text }} {{ row.reason }}
No rejected rows found
Every non-empty source row is valid under the current maximum score and parser settings.
 
{{ formattedJson }}
Tags: Education, Productivity
Test Curve Calculator Curve raw test scores, compare class-average and grade-band shifts, and audit caps, rounding, skipped rows, and formulas before posting grades.
Grade Calculator Calculate a current course grade from points, percentages, or letters, then test weights, dropped scores, late penalties, and remaining-work targets.
Test Score Calculator Calculate a test score from correct, missed, or earned points, then compare percent, grade band, next-band gap, and charted cutoffs.
Weighted Rubric Score Calculator Calculate weighted rubric scores from criterion levels, percent or point weights, grade bands, and audit checks for contribution and setup gaps.
Customize
Advanced
:

A score list can look settled in a spreadsheet while still hiding the pattern that matters. Ten marks can average to a passing result even when several learners sit below the cut score. Another class can share the same average but have a tight middle, a long low tail, or a few high marks pulling the number upward. A score distribution keeps those differences visible by showing where each mark lands on the same 0 to 100 percent scale.

Distribution review matters whenever marks guide feedback, certification, placement, moderation, or follow-up teaching. In a classroom, the spread can show whether one topic needs reteaching. In a training cohort, it can reveal whether a pass threshold is separating a small group from the rest or whether nearly everyone is clustered around the cutoff. In rubric scoring, it helps reviewers see whether marks are concentrated, scattered, or pulled by one part of the range.

The main vocabulary is simple, but each term answers a different question. The mean is the arithmetic average. The median is the middle score after sorting. Quartiles divide the sorted scores into lower, middle, and upper parts. The interquartile range, or IQR, covers the middle 50 percent. Standard deviation describes spread around the mean, while a histogram counts how many scores fall into each score range.

Score distribution reading map Histogram-style score ranges with mean, median, pass threshold, quartile band, and outlier markers. 0% 50% 100% mean median pass line middle 50% outliers

Good score review separates the statistical question from the policy question. Statistics can show that many marks are near a threshold, that the low tail is unusually long, or that the middle 50 percent is narrow. They do not decide whether a test was fair, whether a curve should be applied, or whether a learner mastered a topic. Those decisions need the assessment design, item difficulty, marking notes, attendance context, and any accommodations that affected the results.

Several common mistakes lead to poor score interpretation. Reporting only the mean hides tails and clusters. Treating every outlier as an error can remove a real high or low performance. Comparing two groups without the same maximum score, grade bands, and pass threshold can make ordinary scaling differences look like performance differences. Leaving invalid rows mixed into the data can be worse because a single nonnumeric or out-of-range value can distort the whole summary.

The safest summary combines center, spread, threshold counts, and row-quality checks. That gives enough evidence to describe the group honestly while still leaving room for professional judgment about grading policy, training follow-up, or data cleanup.

How to Use This Tool:

Start with the score rows, set the scale and policy thresholds, then check rejected rows before relying on the class summary.

  1. Paste one score per line into Score data, drop a CSV or TXT file onto the textarea, or use Browse CSV. Recognized score columns include score, mark, marks, points, grade, percent, and percentage.
  2. Set Maximum score to the points possible. Use 100 when your pasted values are already percentages; use the real total, such as 40 or 50, when the rows contain raw marks.
  3. Edit Grade bands as one label and minimum percentage per line, such as A,90, B,80, and F,0. The Band threshold preview shows the parsed ranges.
  4. Set Pass threshold from 0 to 100 percent. The pass-rate badge and Class Snapshot use scores greater than or equal to that cutoff.
  5. Choose an Outlier method. None suppresses review flags, Z-score distance uses the selected standard-deviation threshold, and IQR fence uses quartile fences.
  6. Open Advanced if you need a different Histogram bin size, Decimal places, Show labels setting, or Header preview setting.
  7. Review Class Snapshot, Distribution Buckets, Grade Bands, Score Rows, and Rejected Rows. If validation errors appear or Rejected Rows lists bad input, fix the source rows and rerun the summary before sharing results.
Show labels is useful for internal row review, but keep it off for ordinary summaries. Before copying or sharing row-level material, inspect the output and remove names or labels that should not travel with the score report.

Interpreting Results:

Class Snapshot uses valid rows only. Rows with no numeric score, a negative score, or a score above the selected maximum are held out in Rejected Rows so they do not affect the mean, median, quartiles, pass rate, charts, or outlier count.

Result view Read first Check before deciding
Class Snapshot Valid scores, mean, median, standard deviation, quartiles, shape, top band, pass rate, outliers, and rejected rows. Compare mean with median and look at rejected rows before describing the group with one average.
Distribution Buckets Count, share, and cumulative share for each score range. Look for clusters near the pass threshold or a grade boundary.
Grade Bands Each band minimum, row count, and share under the current band list. Confirm that the band rules match the policy you intend to apply.
Score Rows Raw score, percent, band, and review flag for each valid row. Inspect near-threshold and outlier rows before changing grades or removing records.
Score Spread and Band Balance Bar charts for score buckets and grade-band counts. Use the charts to notice concentration, gaps, and tails, then verify the table values.

Shape labels such as tight cluster, low-tail skew, high-tail skew, wide spread, small sample, and balanced spread are descriptive cues. A high pass rate does not prove that the assessment was easy, and an outlier flag does not prove that a score is wrong. Treat the labels as prompts for review, not as grading decisions.

Technical Details:

Score-distribution statistics are calculated after raw marks are converted to percentages. Converting first makes a 32 out of 40, an 80 out of 100, and a 24 out of 30 comparable on the same 0 to 100 scale. The maximum score therefore controls both the percentage conversion and the validity check for rows that exceed the possible mark.

Center and spread answer different parts of the distribution. The mean uses every valid percentage and moves toward long tails. The median depends on rank order, so it is less affected by extreme values. Quartiles split the sorted values into lower and upper parts, and the interquartile range describes the width of the middle 50 percent. Standard deviation measures how far scores sit from the mean under the current valid set.

Formula Core

The main calculations use valid percentage rows. The standard deviation shown here is the population spread for the current set of valid rows, not a sample estimate for a larger population.

pi = rawimaximum score × 100 mean = ∑i=1n pi n standard deviation = ∑i=1n (pi-mean)2 n pass rate = count of pi≥pass threshold n × 100

For example, a raw mark of 32 with a maximum score of 40 becomes 80 percent. If seven valid percentages sum to 547, the mean is 547 divided by 7, or 78.1 percent when one decimal place is selected. With a 60 percent pass threshold, six passing rows out of seven produce an 85.7 percent pass rate.

Statistic Rules

Statistic Rule Interpretation note
Median Middle sorted percentage; for an even count, the two middle percentages are averaged. Useful when high or low tails make the mean less representative.
Q1 and Q3 Linear interpolation at sorted positions (n - 1) * 0.25 and (n - 1) * 0.75. Q1 to Q3 is the middle 50 percent of valid scores.
Mode Most repeated percentage after grouping values to three decimal places; ties prefer the lower grouped value. Can be unhelpful when nearly every score is unique.
Grade band Each percentage is assigned to the first band whose minimum is less than or equal to that percentage. Changing the band list can change band counts without changing any score.
Histogram bucket Percentages are counted using the selected bin size from 0 to 100; the 100 percent endpoint belongs to the final bucket. Smaller bins show more detail but can make small samples look noisy.

Outlier Rules

Z-score mode standardizes each percentage by its distance from the mean. A row is flagged when the absolute Z-score is greater than or equal to the selected threshold.

zi = pi-meanstandard deviation , flag when | zi | ≥ Z threshold

IQR mode needs at least four valid scores and a nonzero interquartile range. Scores outside the inner fences are flagged; scores exactly on a fence are not outside it.

IQR = Q3-Q1 lower fence = Q1-1.5×IQR upper fence = Q3+1.5×IQR

Shape Cues

Cue Rule Meaning
Small sample Fewer than 5 valid scores. Use statistics as a quick check, not a stable shape claim.
Tight cluster Mean and median are within 2 percentage points, IQR is at most 12 points, and standard deviation is at most 10 points. The average is a fair short summary because scores sit close together.
Low-tail skew or high-tail skew Mean-median gap reaches 4 points in either direction, or skewness cue crosses -0.45 or 0.45. A tail is pulling the mean away from the middle score.
Wide spread IQR is at least 22 points or standard deviation is at least 16 points. Band counts and quartiles describe the group better than one average.
Balanced spread No earlier shape rule is triggered. Center and spread are broadly aligned for a compact summary.

Limitations and Privacy Notes:

Score summaries describe the rows currently provided. They do not infer missing learners, apply item weights, test whether an assessment was fair, or recommend a curve. Files are read in the browser for parsing, and no server-side scoring service is needed for the calculations, but privacy still depends on what you paste, display, copy, and export.

  • Use a CSV or TXT score file smaller than 1 MB when loading local files.
  • Keep Show labels off for anonymous summaries, and inspect copied rows or downloads before sharing.
  • Use the same maximum score, pass threshold, grade bands, bin size, and outlier method when comparing sections or cohorts.

Worked Examples:

Seven score rows

A teacher uses the sample rows with Maximum score at 100 and Pass threshold at 60. Class Snapshot reports 7 valid scores, a Mean of 78.1%, a Median of 83.0%, and a Pass rate of 85.7%. Grade Bands places the rows into A, B, C, D, and F under the default thresholds.

Raw marks out of 40

A training lead pastes 32, 28, and 24 with Maximum score set to 40. Score Rows shows 80.0%, 70.0%, and 60.0%. With Pass threshold at 70, the Pass rate is 66.7% because 80.0% and 70.0% meet the cutoff.

Rejected import cleanup

A CSV contains Maya,105, Noor,absent, and Iris,-3 while Maximum score is 100. Rejected Rows lists the out-of-range, nonnumeric, and negative-score reasons, and Class Snapshot excludes those rows until the source data is corrected.

Advanced Tips:

  • Keep Maximum score fixed when comparing groups. Changing it converts every raw mark to a new percentage and can change every downstream statistic.
  • Use the same Grade bands and Pass threshold for cohort comparisons; otherwise band shares and pass rates may reflect policy settings rather than performance.
  • Choose IQR outliers for a rank-based review of skewed score sets, and use Z-score outliers only when distance from the mean is the question you want to ask.
  • Use a wider Histogram bin size for small classes so one or two rows do not create a misleadingly detailed shape.
  • Turn Header preview on when importing unfamiliar CSV files so the detected score column can be checked before the summary is shared.

FAQ:

Should I trust the mean or the median?

Use both when the result matters. The mean shows the arithmetic average, while the median shows the middle score. A large difference between them means tails, clusters, or outliers need review.

Why did my CSV row get rejected?

Rejected Rows lists rows with no numeric score column, a nonnumeric score, a negative score, or a raw score greater than the selected Maximum score.

What changes when I edit Maximum score?

Every valid raw score is converted again as a percentage. Mean, median, quartiles, pass rate, grade bands, outlier flags, rejected rows, charts, and JSON can all change.

Does an outlier flag mean the score is wrong?

No. It means the score is unusual under the selected outlier method. Check the raw score, maximum score, grading notes, and row context before changing or excluding it.

Can I compare two different classes?

Yes, but keep Maximum score, Grade bands, Pass threshold, Histogram bin size, Decimal places, and Outlier method consistent. Otherwise the differences may come from settings instead of scores.

Glossary:

Mean
The arithmetic average of valid percentage scores.
Median
The middle valid percentage after the scores are sorted.
Quartile
A cut point that divides sorted scores into quarters.
IQR
The interquartile range, equal to Q3 minus Q1.
Standard deviation
A measure of spread around the mean for the valid score set.
Z-score
A score's distance from the mean measured in standard deviations.
Pass threshold
The selected percentage cutoff used for pass-rate calculations.

References: