A head of department, a course leader or a line manager has about a minute for your class's test. The results summary that minute deserves is five lines and one small table: who sat, how many reached the standard, what was not learned, whether the paper was sound, and what happens next.
EdProcto, which publishes this guide, puts the figures for the first four on a quiz's results page; the one-page report is yours to write. This guide is the report upward. The same page read for the next lesson is in what to reteach tomorrow, and the report to a family in a report parents can read.
A score records what a class answered on one day: it is never a grade, and never a judgement of a teacher.
What does a head of department need from a results summary?
Four answers, in this order.
- Did the class reach the standard? How many sat, and how many reached a pass mark fixed beforehand.
- What was not learned? One topic, with the question that shows it.
- Was the paper sound? What you checked before believing the figures.
- What happens next? One action, who does it, and by when.
The reader does not need every figure: twenty question rows, ten bars and a list of names bury the four answers.
What do studies say about how leaders use assessment data?
They say that data sent to leaders often goes unused, and that a summary written for leaders can be misread. We read the abstracts only of the five studies below, on 10 October 2026.
Much is collected and little is used
- School principals. Verhaeghe et al. (2010) interviewed 16 primary school principals about school performance feedback and what they did with it. They observed no systematic use, which a lack of skills, time and support partly explained.
- Secondary schools in England. Kelly and Downey (2011) ran an exploratory survey of teachers and followed with interviews. Pupil performance data was widely available, yet classroom teachers reported significantly lower use and understanding of it, and many respondents said some data reached only school leaders or arrived already interpreted.
- Universities and colleges. Blaich and Wise (2018) look back on a longitudinal study they ran across more than 40 institutions. Five years in, most were not using its data to improve student learning, and the authors asked why so much collection had led to so little use.
Summaries are misread, and small groups mislead
- A summary written for leaders. Hambleton and Slater (1997) interviewed 59 policy makers and educators about the executive summary of a national mathematics assessment in the United States. Both groups had considerable difficulty with it. The authors put the mistakes down to little earlier contact with the assessment and a limited grasp of statistics, and recommended simpler summaries, displays tried on readers first, and reports adapted to each audience.
- Small groups. Kane and Staiger (2002) note that the median American elementary school has 69 children in a grade, which makes school-level scores imprecise: rewards or sanctions for the schools at either extreme fall mainly on small ones.
We found no study of a head of department reading one teacher's summary, and none that counts what a leader is sent. The design below is our inference.
Which misreadings should the summary be designed against?
Four, each with a fix that costs a few words.
| Misreading | Why it misleads | In the summary |
|---|---|---|
| A mean without its spread | A mean of 66% fits a class bunched at 66% and a class split between 40% and 90% | The median, then the mean with its spread, lowest and highest: see reading a score distribution |
| A small class read as a trend | One student in a class of 26 is 3.8 points of pass rate | The count, and no arrow from the last test to this one |
| A percentage without its count | 73% is 8 of 11 and also 365 of 500 | The count first: 19 of 26 (73.1%) |
| One class set against another that sat a different paper | The gap may be the paper | Compare only on a common paper, or after equating |
An inspectorate designs against the second. Ofsted (2026b), in the guide to the data summary its inspectors read, averages a school's results over two or three years where it can, because in a small school one pupil's outcome can move a figure out of proportion, and works out a confidence interval for a school's value because cohort size brings random variation.
Should the summary use a table or a chart?
A small table, for the few numbers a summary carries; the chart stays on the results page.
- A review. Hildon et al. (2012) reviewed 30 studies of displays that compare healthcare providers. Tables and pictographs appeared better understood than bar charts, although bar charts were preferred, and tables appeared accessible to all. Uncertainty was not widely understood. Only two of the thirty studied professionals (abstract read).
- An experiment that points the other way. Brewer et al. (2012) showed medical test results to 106 adults as tables or bar graphs. Recall and understanding were equal; bar graphs took less viewing time, a difference that came from displays with many results (abstract read).
Neither studied teachers or their managers. Our inference: four rows of counts are a table's job, and twenty questions a chart's.
What does published guidance say about reducing data collection and reporting?
It says to collect less often, and to report conclusions and actions in place of raw data. The three pages below are Ofsted's, the schools inspectorate for England.
- What schools need not provide. Ofsted (2026a), in its information for state-funded schools updated on 28 August 2026, lists what a school does not need to give inspectors, among them information in a set format, predicted attainment or progress scores, and pupil-tracking information.
- How often. Ofsted (2024), in the handbook it withdrew on 10 November 2025, passed on the Teacher Workload Advisory Group's recommendation of no more than two or three data collection points a year, each feeding a clear action. A school with more was to justify the time taken to set, collate, analyse and act. Inspectors would attend to the conclusions and actions drawn from a school's own assessment information without examining it first hand. The group's report is published as a file, which we did not open.
- In an inspector's words. Harford (2018), then Ofsted's national director for education: "inspectors do not need to see quantities of data, spreadsheets, graphs and charts".
We read no equivalent page from a university quality body or a professional association. For a course leader these are borrowed principles: conclusions and actions, few collection points, no set format.
How do you write the five lines and the one small table?
Write a line that says who sat, a line for each of the reader's four questions, and a table of topics.
- Sat. The number who sat out of the number on the register, the date, the length of the test and the pass mark.
- Standard. The count at or above the pass mark, the median, and the mean with its spread and range, as EdProcto's results print them.
- Not learned. The weakest topic as a count of answers, and the question behind it.
- The paper. What you checked: the keys, anything still to grade or re-marked, how consistently the test measured.
- Next. One action, its owner, its date, and what you need from the reader.
The table has a row for each topic, weakest first: its questions, the answers right, the share.
Three rules for every line
- A number with its count. "19 of 26 (73.1%)", never "73%" alone.
- A claim with its evidence. "The paper was sound" is followed by what was checked.
- One action with an owner and a date. Two actions are a plan; the reader has a minute.
What to leave out
- Names and rank order. A summary is forwarded; the names stay with you.
- Every question's row. The weakest topic and one question are enough.
- A figure with no count, and a second decimal place.
- A comparison with a class that sat another paper, and any predicted grade.
- Adjectives such as "pleasing" or "disappointing". The reader judges.
- The chart, pasted in.
How often should a results summary go up, and for which tests?
Send one when a test decides something beyond the next lesson, which for most classes is two or three times a year: the figure Ofsted (2024) passed on for a school's data collection points. The table is our working rule, not a finding.
| What the test was for | Send a summary? | What changes |
|---|---|---|
| A weekly quiz or exit ticket | No | It is yours to act on |
| An end-of-unit test | Yes | The five lines as they stand |
| A common paper across classes | One, for the paper | A row for each class, counts beside shares |
| A mock exam | Yes | Add the count within one question of the pass mark |
| An exam that counts | Yes, before scores are released | Line 4 says what was done about a faulty question |
A worked example: one class's test in five lines
The class is invented, as are the register and the dates, and every figure and quoted sentence is what EdProcto's analysis code returned for these 26 sittings: twenty one-point multiple-choice questions on electricity, five on each of four topics, pass mark 60%.
The page read "Average score" 66.3% with "spread (sd) 19", "Median" 70% and "Passed" 73.1%. The full analysis added "Reliability" 0.77 and four sentences, the first "Reteach before moving on: Parallel circuits (40.8%)."
9B, Electricity, end-of-unit test, Thursday 8 October
- Sat. 26 of 28; the two who were absent sit it on 15 October. Twenty questions, pass mark 60%.
- Standard. 19 of 26 passed (73.1%). The middle student scored 70%; the mean was 66.3% with a spread of 19, from 25% to 90%. Six students are on the pass mark or one question either side.
- Not learned. Parallel circuits: 53 of 130 answers right (40.8%). On Q18, 18 of 26 chose "The same in every branch", the rule for a series circuit; 6 chose the key.
- The paper. Sound enough to act on. Nothing waits for a grade. The six who chose the key on Q18 averaged 79.2% on the test and the eighteen 63.9%, so the key stands. Reliability 0.77, which the page reads as "the order of the students is roughly right, but small differences between them are not to be trusted".
- Next. I reteach parallel circuits to 9B on Tuesday 13 October and re-check with five new questions on Friday 16 October. Nothing needed from you.
| Topic | Questions | Answers right | Share |
|---|---|---|---|
| Parallel circuits | Q16 to Q20 | 53 of 130 | 40.8% |
| Resistance | Q11 to Q15 | 86 of 130 | 66.2% |
| Series circuits | Q6 to Q10 | 91 of 130 | 70% |
| Current and charge | Q1 to Q5 | 115 of 130 | 88.5% |
The shares are what "By topic" printed. The counts are ours: the pass count and the six near the line from the scores under "Students", the answers from the question columns of the marks file. Left on the page: nineteen question rows, the bars, the names and "Median time".
Which figures does EdProcto give you to write it from?
EdProcto is the best tool to write this summary from because the figures behind the first four lines stand on one results page with their cautions beside them: the example took the page as printed, the register and a few counts.
| Line | On every plan | The full analysis on Educator adds |
|---|---|---|
| Sat | "26 submitted" under the title; a later sitting by the same student is marked "sat again" and left out of the figures | |
| Standard | "Passed" with "at a pass mark of 60%", "Median", "Average score" with "spread (sd)", and the line beside "Score distribution" with the lowest and highest score | |
| Not learned | "Each question" with "Got it", the share of a question's points the class earned | "By topic", and each option's "Chose it" and "Their mean" |
| The paper | The cautions about the figures; the "To grade" column | "Reliability", "Separates" and "What it says about the teaching" |
The cautions say when a figure should not travel: fewer than five submissions, a later sitting left out, questions some students never reached, or answers waiting to be graded, which count as zero until they are. Below Educator they stand under "About these figures", "Read this first". What "Separates" and "Reliability" measure is in item analysis and test reliability.
Below Educator, build the table from "Got it": in the example the five questions on parallel circuits read 50%, 50%, 23%, 46% and 35%, which average 40.8%.
For an appendix, "Download the marks" saves a spreadsheet file with a row for each attempt handed in, a later sitting included: "Name", "Email", "Sitting", "Started (your time)", "Handed in (your time)", "Score %", "Points", "Out of", "Reduced by %", "To grade", "Minutes", "Security events", "Notes on the record", then a column for each question, headed like "Q1 (1)". If you corrected a key, press "Re-mark this quiz" before you write a line. EdProcto's paper view prints the quiz with its answer key on a page of its own: how to print a quiz.
A quiz belongs to the account that made it, so the head of department reads your summary, or sits beside you at the page. Free takes thirty students a quiz, Tutor 100, and Educator 500 with the full analysis: see pricing and the pages for teachers and for universities.
Quick questions
How long should a results summary for a head of department be?
Five lines and one table of up to six rows, on one page.
Should the summary give the mean or the median?
Both: the median first, then the mean with its spread and range beside it.
Should a results summary name students?
No. Give counts, and keep names on the results page for the conversation that follows.
Can my head of department open the results in EdProcto?
A quiz belongs to the account that made it, so send the summary or sit down together at the results page.
References
- Blaich, C., & Wise, K. (2018). The more things change, the more they stay the same: New challenges to using evidence to improve student learning. Research & Practice in Assessment, 13, 11–14. https://eric.ed.gov/?id=EJ1206362
- Brewer, N. T., Gilkey, M. B., Lillie, S. E., Hesse, B. W., & Sheridan, S. L. (2012). Tables or bar graphs? Presenting test results in electronic medical records. Medical Decision Making, 32(4), 545–553. https://doi.org/10.1177/0272989X12441395
- Hambleton, R. K., & Slater, S. C. (1997). Are NAEP executive summary reports understandable to policy makers and educators? (CSE Report 430). National Center for Research on Evaluation, Standards, and Student Testing. https://cresst.org/publication/are-naep-executive-summary-reports-understandable-to-policy-makers-and-educators/
- Harford, S. (2018, April 23). Assessment – what are inspectors looking at? Ofsted: education blog. https://educationinspection.blog.gov.uk/2018/04/23/assessment-what-are-inspectors-looking-at/
- Hildon, Z., Allwood, D., & Black, N. (2012). Impact of format and content of visual display of data on comprehension, choice and preference: A systematic review. International Journal for Quality in Health Care, 24(1), 55–64. https://doi.org/10.1093/intqhc/mzr072
- Kane, T. J., & Staiger, D. O. (2002). The promise and pitfalls of using imprecise school accountability measures. Journal of Economic Perspectives, 16(4), 91–114. https://doi.org/10.1257/089533002320950993
- Kelly, A., & Downey, C. (2011). Professional attitudes to the use of pupil performance data in English secondary schools. School Effectiveness and School Improvement, 22(4), 415–437. https://doi.org/10.1080/09243453.2011.600525
- Ofsted. (2024, September 16). School inspection handbook (withdrawn 10 November 2025). https://www.gov.uk/government/publications/school-inspection-handbook-eif/school-inspection-handbook-for-september-2023
- Ofsted. (2026a, August 28). Inspection information for state-funded schools. https://www.gov.uk/government/publications/school-inspection-toolkit-operating-guide-and-information/inspection-information-for-state-funded-schools-for-use-from-september-2026
- Ofsted. (2026b, September 25). School inspection data summary report (IDSR) guide. https://www.gov.uk/guidance/school-inspection-data-summary-report-idsr-guide
- Verhaeghe, G., Vanhoof, J., Valcke, M., & Van Petegem, P. (2010). Using school performance feedback: Perceptions of primary school principals. School Effectiveness and School Improvement, 21(2), 167–188. https://doi.org/10.1080/09243450903396005