Two scores are equivalent only if they measure the same thing in the same way. A practice score, a teacher’s mark, a school test and a real result usually differ in task, length and marking. Scores from learners on 0510 and 0511 carry an extra difference, because speaking is treated differently on each route.
Comparing unlike scores leads to false comfort or false worry. This lesson gives you a short test to run before you compare anything. It follows the ideas in speaking endorsement and count-in speaking.
What makes scores “unlike”?
Ask five questions about each score:
- Which code was the learner entered for?
- Which task was it: reading, listening, writing or speaking?
- How long was the task and how many marks was it out of?
- Who marked it, and with which mark scheme?
- Which year’s syllabus was the task written for?
If any answer differs, the scores are not equivalent. They may still be useful, but for different reasons.
How to compare fairly, step by step
- Write both scores as the marks given, such as 34 out of 40.
- Write the task and the marker beside each.
- If anything differs, do not compare the scores. Compare the skill comments instead.
- If everything matches, convert to percentages for a rough comparison.
- Track yourself over time on one task type with the same marking.
Worked example
Aisha is on 0510 and Kavin is on 0511. Aisha scored 34 out of 40 on a reading practice from a worksheet. Kavin scored 27 out of 30 on a reading quiz written by his teacher.
As percentages: Aisha 34 ÷ 40 = 0.85, so 85%. Kavin 27 ÷ 30 = 0.90, so 90%.
Kavin seems five points ahead. Now run the five questions. The codes differ, the tasks differ in length, the markers differ and nothing is known about the syllabus year.
So the two percentages are not equivalent, and Aisha should not conclude she is behind.
A better use of the numbers: Aisha compares her 85% with her own next reading practice from the same source, and Kavin does the same with his.
The mistake to watch for
Mistaken reasoning: “Kavin got 90% and I got 85%, so he is better at reading than I am.”
The percentages look alike, but they came from different tasks and markers.
The correction is to hold back judgement until the five questions match. Even then, a single score is one data point, not a level.
Check yourself
1. Sam scored 18 out of 20 in one quiz, and Lin scored 40 out of 50 in another. Which percentage is higher, and can Sam and Lin be compared?
Show answer
Sam: 18 ÷ 20 = 0.90, so 90%. Lin: 40 ÷ 50 = 0.80, so 80%. Sam’s percentage is higher, but the quizzes are different, so they cannot be compared without knowing the tasks and marking.
2. Name two of the five questions to ask before comparing scores.
Show answer
Any two of: which code, which task, how long and how many marks, who marked it, which syllabus year.
3. What is a fair way to use your own score?
Show answer
Compare it with your own later scores on the same kind of task, marked the same way, and note which skill changed.
Where this leads next
Put the whole module together with the 0510 and 0511 route awareness practice set. Keep your results in the mistake log and retest queue, and use the summary selection studio and writing revision compare view to measure change on the same task.
A teacher in online one-to-one English as a Second Language tuition can set one measure for your progress and review it with you.