Edexcel GCSE statistics 2017 unofficial markscheme

•

a) 8 } For these 2 it was just reading off the graph [1]

•

b) 28 } and doing 87 – 59 = 28 (or something like that) [2]

•

c) No, as there may be a different number of people in the UK and USA and we only know the % rather than actual numbers. Trends may also differ significantly for each country. [1]

•

d) IQR and Median were higher for tablet users, suggesting that overall the average age of a smartphone user is younger than for a tablet user (in the USA). Plus reference to the values and the difference. [4]

•

- 35-56 year olds - 10% more guessed the price of a stamp correct, but both those and the 16-34 guessed too low(?) similarly. Plus evidence in the form of values/percentage. [2]

•

a) Electoral roll/register [1]

•

b) Is cheaper as no postage costs for researcher or respondent, so higher response rate. [1]

•

c) One source of bias – interviewers may choose people they know to interview [1]

•

d) 80% said they would vote [2]

•

e) Method 1 = cluster sampling and Method 2 = stratified sampling [2]

•

f) Method two is better because 1) you will not get too many of a certain type of person (i.e. males) 2) less chance of bias [2]

•

a) The box plot can be used to support this as it is slightly negatively skewed, and the median value shows that over 50% paid over 200k for their house [1]

•

b) 1.5 x IQR + UQ = 455,000. 500,000 is higher than the upper outlier boundary so is an outlier [3]

•

a) Age, as it is quantitative (which one can be shown in a stem&leaf) [2]

•

b) Time taken, as it will be discrete and has suitable class widths(?) [2]

•

c) Median (The name of the statistic needed) [1]

•

a) Why he would have to collect primary data: he is the first to carry out this study [1]

•

b) If the sample is too small it will not be representative [1]

•

c) Tick boxes with non-overlapping values (I did less than 0.5 miles, 0.5-1 mile, 1.1-1.5 miles, 1.6-2 miles, more than 2 miles) [2]

•

a) Show the sum of d^2 is 10: 0+0+1+4+4+1+0 = 10 [1]

•

b) SRCC = 0.82 [3]

•

c) Relatively strong correlation [1]

•

d) Yes, because the positive correlation suggests that the higher ice cream consumption has a relationship with the GDP value thing. [2]

•

a) At first the prices increased rapidly, then they slowly fell by 5% by 2015. [1]

•

b) 2011 price = £3304.88 (3032 + 9% increase). [1]

•

c) Find the price for 2015. Show that if it increased by 25% it would be (something like) £3720 and the price for 2015 is £3820 so more than a 25% increase. [2]

•

a) 19/30 [1]

•

b) 16/30 [2]

•

c) 1/25 chance [2]

•

d) No, Greg is wrong. Chance of over 29 degrees was 10/30, chance of over 29 when rain was 6/30. So more likely to be over 29 when NOT raining. [3]

•

a) I got 500 but I think the answer was actually 402. [2]

•

b) It allowed the geese to move around as they would naturally, to distribute themselves. [1]

•

c) No, because the population may have increased more than he thought, he only took one sample. More offspring may have been born, geese could have died or moved to live somewhere else. [2]

•

a) Mean = 51 (based on the scores I think it gave). I can’t remember his scores, I think it was 55, 41 and 57? But you had to find the mean. [1]

•

b) Weighted mean will be less because each component counts for a different % so one holding more weight counts for the most amount of mark, for example. [1]

•

c) Weighted mean = (0.35 x 55) + (0.45 x 41) + (0.25 x 57). I don’t know the exact marks or weightings, sorry, but I THINK that is the formula. [2]

•

d) Number of students higher than him – I got something like 28. I just used the graph to add up the frequencies and then halved the class width that he was in and added that lol. [2]

•

e) 11.5 (or something like that was what I got for the Standard deviation.) [2]

•

f) SD is better because it will show how the data is spread and how close to the mean average it is. [1]

•

a) 0.7 chance of red [1]

•

b) 0.49 chance of red both times [2]

•

c) 0.83. I can’t remember exactly what it askes, but it was something like more than ½ of 5? This wording was confusing but I did p^5, 5p^4q, 10p^3q^2. Which would be 0.16 + 0.36 + 0.31 = 0.83 [3]

•

d) Show that the most likely number is 4 = 5p^4q worked out is 0.36 to 1dp = 0.4 which is 4 marbles [3]

•

e) No, binomial can only be used for independent trials. [2]

•

a) 0.6 (plus working out) [2]

•

b) She did better in the vault, as her standardised score was higher for that than the beam. Evidence in the fact that -0.6 < 0.5 [2]

•

c) Could it be modelled by a normal distribution? NO, as 95% should lie between 13.3 and 15.7 and the highest score at 15.3 is below that 95%. Plus working to show mean +- 2 sd. [4]

•

a) 8 } For these 2 it was just reading off the graph [1]

•

b) 28 } and doing 87 – 59 = 28 (or something like that) [2]

•

c) No, as there may be a different number of people in the UK and USA and we only know the % rather than actual numbers. Trends may also differ significantly for each country. [1]

•

d) IQR and Median were higher for tablet users, suggesting that overall the average age of a smartphone user is younger than for a tablet user (in the USA). Plus reference to the values and the difference. [4]

•

- 35-56 year olds - 10% more guessed the price of a stamp correct, but both those and the 16-34 guessed too low(?) similarly. Plus evidence in the form of values/percentage. [2]

•

a) Electoral roll/register [1]

•

b) Is cheaper as no postage costs for researcher or respondent, so higher response rate. [1]

•

c) One source of bias – interviewers may choose people they know to interview [1]

•

d) 80% said they would vote [2]

•

e) Method 1 = cluster sampling and Method 2 = stratified sampling [2]

•

f) Method two is better because 1) you will not get too many of a certain type of person (i.e. males) 2) less chance of bias [2]

•

a) The box plot can be used to support this as it is slightly negatively skewed, and the median value shows that over 50% paid over 200k for their house [1]

•

b) 1.5 x IQR + UQ = 455,000. 500,000 is higher than the upper outlier boundary so is an outlier [3]

•

a) Age, as it is quantitative (which one can be shown in a stem&leaf) [2]

•

b) Time taken, as it will be discrete and has suitable class widths(?) [2]

•

c) Median (The name of the statistic needed) [1]

•

a) Why he would have to collect primary data: he is the first to carry out this study [1]

•

b) If the sample is too small it will not be representative [1]

•

c) Tick boxes with non-overlapping values (I did less than 0.5 miles, 0.5-1 mile, 1.1-1.5 miles, 1.6-2 miles, more than 2 miles) [2]

•

a) Show the sum of d^2 is 10: 0+0+1+4+4+1+0 = 10 [1]

•

b) SRCC = 0.82 [3]

•

c) Relatively strong correlation [1]

•

d) Yes, because the positive correlation suggests that the higher ice cream consumption has a relationship with the GDP value thing. [2]

•

a) At first the prices increased rapidly, then they slowly fell by 5% by 2015. [1]

•

b) 2011 price = £3304.88 (3032 + 9% increase). [1]

•

c) Find the price for 2015. Show that if it increased by 25% it would be (something like) £3720 and the price for 2015 is £3820 so more than a 25% increase. [2]

•

a) 19/30 [1]

•

b) 16/30 [2]

•

c) 1/25 chance [2]

•

d) No, Greg is wrong. Chance of over 29 degrees was 10/30, chance of over 29 when rain was 6/30. So more likely to be over 29 when NOT raining. [3]

•

a) I got 500 but I think the answer was actually 402. [2]

•

b) It allowed the geese to move around as they would naturally, to distribute themselves. [1]

•

c) No, because the population may have increased more than he thought, he only took one sample. More offspring may have been born, geese could have died or moved to live somewhere else. [2]

•

a) Mean = 51 (based on the scores I think it gave). I can’t remember his scores, I think it was 55, 41 and 57? But you had to find the mean. [1]

•

b) Weighted mean will be less because each component counts for a different % so one holding more weight counts for the most amount of mark, for example. [1]

•

c) Weighted mean = (0.35 x 55) + (0.45 x 41) + (0.25 x 57). I don’t know the exact marks or weightings, sorry, but I THINK that is the formula. [2]

•

d) Number of students higher than him – I got something like 28. I just used the graph to add up the frequencies and then halved the class width that he was in and added that lol. [2]

•

e) 11.5 (or something like that was what I got for the Standard deviation.) [2]

•

f) SD is better because it will show how the data is spread and how close to the mean average it is. [1]

•

a) 0.7 chance of red [1]

•

b) 0.49 chance of red both times [2]

•

c) 0.83. I can’t remember exactly what it askes, but it was something like more than ½ of 5? This wording was confusing but I did p^5, 5p^4q, 10p^3q^2. Which would be 0.16 + 0.36 + 0.31 = 0.83 [3]

•

d) Show that the most likely number is 4 = 5p^4q worked out is 0.36 to 1dp = 0.4 which is 4 marbles [3]

•

e) No, binomial can only be used for independent trials. [2]

•

a) 0.6 (plus working out) [2]

•

b) She did better in the vault, as her standardised score was higher for that than the beam. Evidence in the fact that -0.6 < 0.5 [2]

•

c) Could it be modelled by a normal distribution? NO, as 95% should lie between 13.3 and 15.7 and the highest score at 15.3 is below that 95%. Plus working to show mean +- 2 sd. [4]