Professional Documents
Culture Documents
Instructions: GEN1008 / MED1018 / GED1008 Mid-Term Test Page 1 of 17
Instructions: GEN1008 / MED1018 / GED1008 Mid-Term Test Page 1 of 17
Instructions: GEN1008 / MED1018 / GED1008 Mid-Term Test Page 1 of 17
INSTRUCTIONS
1. Time allowed: 1.5 hours.
2. Answer all questions on this question-answer paper.
3. The use of HKEA approved calculator(s) is allowed.
4. The use of dictionary and other electronic devices are prohibited.
Question Score
1
2
3
4
5
Total / 50
A health care professional wishes to analyse the Body Mass Index (BMI) of teenagers
aged from 13 to 19. He conducted a survey and collected the BMI readings (in kg/m 2)
of 24 teenagers. The data are shown in the table below. It is given that the mean of
data below is 21.25 kg/m2.
25 20 22 31 20 19 23 25
18 16 20 18 25 17 22 26
18 22 17 19 15 21 28 23
a) Construct a frequency distribution table for the above sample data using 6 classes.
Use 15 as the lower limit of the first class. Write a title and include the class
limits, the class boundaries and the frequency as columns in your table. [4]
b) What is the shape of the distribution of data in the sample? Compare the mean
and the median in the sample data. [2]
c) Given ∑𝑥 2 = 11200, find the standard deviation of the sample data. [2]
d) How will the standard deviation in (c) be affected if the two largest values in the
sample are changed to numbers higher than 32 kg/m2? Explain your answer
briefly. [2]
[Solution]
a)
Class Limits Class Boundaries Frequency
15 - 17 14.5 - 17.5 4
18 - 20 17.5 - 20.5 8
21 - 23 20.5 - 23.5 6
24 - 26 23.5 - 26.5 4
27 - 29 26.5 - 29.5 1
30 - 32 29.5 - 32.5 1
Total 24
Frequency distribution of the Body Mass Index of 24 teenagers in the
sample.
(4 marks)
b) The sample data is positively skewed. The mean is larger than the median.
(2 marks)
d) The standard deviation will be larger since the values in the data set are more
dispersed. (2 marks)
Consider a population of university graduates and their scores in an aptitude test are
analysed. There are two groups of graduates: Group A graduates have degrees in
humanity and Group B graduates have degrees in science. The mean and standard
deviation of scores of both groups and all graduates are shown in the table below. It is
found that the distribution of scores of all graduates follows a bimodal distribution.
[Solution]
So, at least 80% of all graduates will have aptitude scores from 43.89 and 98.9
marks. (3 marks)
d) Yes, this is reasonable, since at least 80% of students have scores higher than
43.89, which is higher than the passing mark. (1 mark)
There is a study on the reading time (hours per week) of secondary school students in
Hong Kong. A recent report reveals that the population mean time is 3.2 hours per
week and the population standard deviation is 0.8 hours per week. It is assumed that
the reading time follows a normal distribution.
a) Find the proportion of secondary school students in Hong Kong whose reading
time is less than 2.4 hours per week. [2]
b) Suppose a researcher wishes to select students whose reading time is in the
middle 50% of the reading time of the population. Find the range of the reading
time of students who may be selected for the research. [4]
c) Another researcher selected a sample of 45 students from the whole population.
What is the probability that the mean reading time of the sample is less than 2.8
hours per week? [2]
d) Explain whether there are any changes in your answer in (c) if the reading time
follows a left-skewed distribution, instead of a normal distribution. [2]
[Solution]
Students whose reading time is between 2.66 and 3.74 hours per week may be
selected for the research.
(4 marks)
This is because, by the Central Limit Theorem, the distribution of sample means
follows a normal distribution when the sample size is large (n > 30) for any
population distribution. (1.5 marks)
Remarks:
- If the student just states “because of the Central Limit Theorem”, deduct 0.5
marks.
- If the student just states “because the sample size is large”, deduct 0.5 marks.
Suppose you work in a research team to investigate whether cancer patients are
satisfied with their lives after receiving a radiation therapy treatment. In a pilot study
(i.e., a smaller scale study), a sample proportion 0.64 is used to determine a 90%
confidence interval, which is [0.5692, 0.7108]. The sample proportion is the
proportion of patients in the sample who are satisfied with their lives.
a) Based on the results from the pilot study, is it reasonable to conclude that more
than half of cancer patients are satisfied with their lives after receiving a
radiation therapy treatment? Explain your answer. [2]
b) What is the margin of error of the confidence interval in the pilot study? [1]
c) Suppose you want to conduct a larger scale study to find the confidence interval
of proportion. The desired level of confidence is 95% and the margin of error is
half of that in the pilot study result. By using the sample proportion in the pilot
study, find the minimum sample size required. [3]
d) You finally used a sample of 750 patients in your study and 492 of them are
satisfied with their lives after receiving the treatment. Find the 95% confidence
interval of the proportion using this sample data. Interpret your answer in the
context of the subject matter. [4]
[Solution]
a) Yes, it is reasonable since the 90% confidence interval contains values that are
larger than 0.5. (2 marks)
𝑝̂ (1 − 𝑝̂ ) 0.656(1 − 0.656)
𝑝̂ ± 𝑧√ = 0.656 ± 1.96√
𝑛 750
We are 95% confident that between 62.2% and 69% of cancer patients are
satisfied with their lives after receiving the treatment. (2 marks)
[Solution]
c) We are 95% confident that the population mean fasting glucose level of all
patients with diabetes is between 5.11 mmol/L and 5.49 mmol/L one week after
the treatment. (2 marks)
d) (i) No change
(ii) The interval is wider (2 marks)
e) 𝐻0 : 𝜇 = 6
𝐻1 : 𝜇 < 6 (claim)
(2 marks)
THE END
Population variance:
2
∑(𝑋 − 𝜇)2 ∑𝑋 2
2
∑𝑋 2
𝜎 = 𝑜𝑟 𝜎 = −( )
𝑁 𝑁 𝑁
Sample variance:
∑(𝑋 − 𝑋̅)2 𝑛(∑𝑋 2 ) − (∑𝑋)2
𝑠2 = 𝑜𝑟 𝑠2 =
𝑛−1 𝑛(𝑛 − 1)
Standard score:
𝑋−𝜇
𝑧=
𝜎
Sample proportion:
𝑥
𝑝̂ =
𝑛
Confidence interval for a proportion:
𝑝̂ (1 − 𝑝̂ )
𝑝̂ ± 𝑧𝛼/2 √
𝑛