MIDTERM SOLUTIONS
Question 1
A) The data will be very discrete and right skewed.
Answer C is not a sufficient reason to call a histogram
non-normal. For instance, heights of adult males must be
positive, but the histogram is quite normal looking.
Question 2
B) It's a t with 51 degrees of freedom and 0.975
Question 3
[2.4-2.009*1.9/sqrt(52),2.4+2.009*1.9/sqrt(52)]
The results are [1.87066,2.9293]
Question 4
Yes.
Question 5
The answers are B) and C). The fact is that everything changes, but a
and d say that the s.d. and the interquartile range remain constant.
Question 6
The answers are B),C),D),E) and F). The only one that's not
appropriate is the paired t-test.
Question 7
The answers are B),C),D),E) and F). Again the only not appropriate is
the paired t-test.
Question 8
The answers are D), E) and F).
Question 9
The answer is C)
Question 10
The answer is A).
Question 11
The answers is the second. The 99% confidence is wider. If you want to
be sure at the 99% you need to take more elements as possible answers,
so the range of possible values will be more spread.
Question 12
It's false. It reflects the population.
Question 13
t*=1.761
distribution centered at 1
chance=0.23
Nearest answer: c=0.19
Question 14
It decreases.
Question 15
The median is about 38
Question 16
The interquartile range is about 20 units.
Question 17
You are supposed to choose the paired t-test, so there's strong
evidence of a difference.
Question 18
p=0.0085
Question 19
paired-t-test
Question 20
.41
Question 21
[.13,.69]
Question 22
0.1 and 0.2 are considered correct answers for this problem.
Question 23
0.1 (2/20)
Question 24
Both a) and b) are considered valid answers for this problem.
Question 25
The answer is 0.0214.