June 2018 Unit 4 Q4
4. Arwyn collects data about household expenditure on food. He records the weekly expenditure on food for 80 randomly selected households from across Wales.
| Cost, \(x\) (£) | \(x<40\) | \(40\leqslant x<50\) | \(50\leqslant x<60\) | \(60\leqslant x<70\) | \(70\leqslant x<80\) | \(80\leqslant x<90\) | \(x\geqslant 90\) |
|---|---|---|---|---|---|---|---|
| Number of households | 5 | 11 | 16 | 18 | 15 | 9 | 6 |
(a) Explain why a normal distribution may be an appropriate model for the weekly expenditure on food for this sample. [1]
Arwyn uses the distribution \(\mathrm{N}(64,\,15^2)\) to model expenditure on food.
(b) Find the number of households in the sample that this model would predict to have weekly food expenditure in the range
(i) \(60\leqslant x<70\),
(ii) \(x\geqslant 90\). [4]
(c) Use your answers to part (b)
(i) to comment on the suitability of this model,
(ii) to explain how Arwyn could improve the model by changing one of its parameters. [2]
(d) Arwyn’s friend Colleen wishes to use the improved model to predict household expenditure on food in Northern Ireland. Comment on this plan. [1]
| Scheme | Marks | Comments |
|---|---|---|
| Valid reason e.g.(Approximately) symmetrical, taller in the middle, tails off at the ends, etc | E1 | Do not accept bell curve alone. |
| Scheme | Marks | Comments |
|---|---|---|
| (i) \(P(60\le X\le 70)=0.26055\) | M1 | M1 for 0.26055 or 80×’their probability’. |
| Predicted number = 21 | A1 | Allow 20.84 or ’20 or 21’ Use of tables gives \(P(60\le X<70)=0.2618\) Predicted number is 21 |
| (ii) \(P(X\ge 90)=0.041518\) | M1 | M1 for 0.041518 or 80×’their probability’. |
| Predicted number = 3 | A1 | Allow 3.32 or ’3 or 4’ Use of tables gives \(P(X\ge 90)=0.0418\) Predicted number is 3 |
| Scheme | Marks | Comments |
|---|---|---|
| (i) Valid comment e.g.18 is smaller than (predicted) 21 and 6 is bigger than (predicted) 3 so may not be the best model. | E1 | Must include reference to model not being ideal/could be improved AND comparison of predicted values and actual values. |
| (ii) Model could be improved by increasing the variance/standard deviation. (This would ‘flatten out’ the curve. It would lower the middle and lift up the tails.) | E1 |
| Scheme | Marks | Comments |
|---|---|---|
| Valid comment. e.g. May not be suitable since the weekly household expenditure on food in Northern Ireland may have a different distribution. e.g. May be suitable as Northern Ireland is part of the UK and has a similar socioeconomic status to Wales. | E1 | |
| [8] |