Worked example · Drama · Year 9
DCF: Data and computational thinking → Data and information literacy
A secure Year 9 response and a full answer key, annotated so you can see which success criterion each part meets. Key answers: total stars 233; mean 3.88; median 4; mode 5; range 4; 70% gave 4 or 5 stars; response rate 17.6%; group chat mean 4.07; QR code mean 3.47; real difference between the nights 0.15 stars. Show this after pupils have tried the tasks themselves.
| Q | What is wrong | How it affects the data or the people answering |
|---|---|---|
| 3 | Leading: “love our amazing songs” tells people what to think. | People are pushed towards a good answer, so the result is not their real opinion. |
| 3 | Every option says “loved”. There is no way to say you did not like the songs. | Anyone who could not hear the songs has to pick a “loved” answer or give up, so the data can only ever look positive. |
| 6 | Asks for a full name, home address and phone number, and makes them required. | The school does not need them to improve a show. Some people will stop and not send the form at all (comment 6), and the school now holds private data it has to keep safe. |
(b) “How did you find the songs?” Options: Very good / Good / OK / Poor / Very poor / I could not hear them.
(c) Remove Question 6, or make it optional and only ask for an email address if people want news of next year’s show.
Each problem is linked to its effect on the data or on the people answering. The new Question 3 is neutral, has a balanced set of options and adds an option the old form could never have recorded. The answer to (c) applies the idea of collecting only the personal data you need. Evidences criterion 1.
(a) 5 × 27 = 135 · 4 × 15 = 60 · 3 × 8 = 24 · 2 × 4 = 8 · 1 × 6 = 6. Total = 135 + 60 + 24 + 8 + 6 = 233 stars.
(b) Mean = 233 ÷ 60 = 3.8833... = 3.88 to 2 decimal places. (c) Mode = 5 stars (27 people). (d) Range = 5 − 1 = 4.
(e) Counting up: 1 star takes values 1 to 6; 2 stars 7 to 10; 3 stars 11 to 18; 4 stars 19 to 33. The 30th and 31st values are both 4, so the median is 4 stars.
(f) 27 + 15 = 42. 42 ÷ 60 × 100 = 70%.
(g) He is using the mean (3.88). The mean is pulled down by the 6 people who gave 1 star. The median (4) is a fairer headline because it is the middle opinion. The mode (5) would be too flattering, because fewer than half the people gave 5 stars.
The working is set out so each step can be checked. The median is found by counting through the frequency table, not by taking the middle of 1 to 5. Part (g) compares the 3 averages and chooses one with a reason. Evidences criterion 2.
(a) 60 ÷ 340 × 100 = 17.647... = 17.6%.
| Group | Working | Mean |
|---|---|---|
| Family group chat (41) | 110 + 36 + 12 + 6 + 3 = 167; 167 ÷ 41 = 4.073... | 4.07 |
| Programme QR code (19) | 25 + 24 + 12 + 2 + 3 = 66; 66 ÷ 19 = 3.473... | 3.47 |
(c) The QR code group is more likely to show what the whole audience thought. First, every audience member got a programme, so anyone could answer, while the group chat only reached the cast’s families. Second, families of the cast are proud of their own children, so they are likely to give higher ratings, which is why their mean is 0.6 stars higher. However, 19 is a small group, so even the QR results are not certain.
(d) People who could not hear may have enjoyed it less and not bothered to scan the code. They are also less likely to be cast families, who usually sit near the front.
Both means are correct with working. The pupil links the way the link was shared to sample bias, compares the 2 groups with a figure, and still notes the limit of a small sample. Evidences criterion 3.
(a) The vertical axis starts at 3.7, not 0, so only the tops of the bars are shown. Friday’s bar looks almost 3 times as tall, but the real difference is 3.94 − 3.79 = 0.15 stars, which is tiny.
(c) “Both nights scored nearly 4 stars from the people who replied.”
The pupil names the cause (the axis does not start at 0), works out the real difference, and draws bars that start at 0 so their heights are fair. The new headline is accurate and admits that not everyone replied. Evidences criterion 4.
Dear Mr Bowen, the data show that most people who answered enjoyed Sinderela: the median rating was 4 stars, 70% gave 4 or 5 stars and 52 out of 60 would come again. However, only 17.6% of the audience answered, and 41 of the 60 came through the cast families’ group chat. Their mean was 4.07, but people using the programme QR code gave only 3.47, so the overall results are probably too generous. The data cannot tell you what the whole audience thought, and Question 3 tells you nothing because every option says “loved”. Next year, put microphones near the singers or speakers at the back, because 2 comments say the back rows could not hear the songs. On the form, ask a fair question about the songs and remove the name and address question.
136 words. It uses 7 accurate figures, separates what the data can and cannot show, and bases the change to the show on evidence from the comments rather than on the flattering Question 3. Both form changes fix problems found in Step 1. Evidences criterion 5.
Thursday: 24 ÷ 150 × 100 = 16%. Friday: 36 ÷ 190 × 100 = 18.9%. Friday had the higher response rate, which makes sense because the group chat link was shared on Friday night.
Question 3 only offers 3 ways of saying “loved”, so the 32 people show how strongly people agreed with a question that told them what to think. Someone who disliked the songs had no answer to choose.