HƯỚNG DẪN ÔN FINAL

1. Ôn tập những nguồn nào?

- Testbank (chỉ để ôn tập; làm trước dò kết quả sau).
- Quizlet (ưu tiên làm những link đề tổng hợp).
- Bài tập cuối sách + bài tập cuối mỗi section trong mỗi chap.

2. MCQs (close book)

- Trong phòng thi được mang tài liệu gì? (có cheatsheet không?)
- Được phát cho công thức không? (công thức được phát bao gồm công thức gì?)
- Được cho appendix nào? (giả sử cho Appendix C1, thì mọi người phải tập xem
appendix C1)
- Casio

3. MCQs (open book)

- Được dùng Excel không? Được dùng Internet không? (để tránh misconduct)
- Vẫn phải học bài kĩ
- Chuẩn bị summary của 12 chapters.
- Excel + Casio

4. MCQs + tự luận
- Ôn tự luận: làm hết tất cả các bài tập cuối sách và cuối mỗi section trong chapter.
- Chú ý cách trình bày bài tự luận, ghi lời giải, công thức rõ ràng.

5. Tự luận
CHAPTER 4
1. Median
a. 1; 3; 5; 6; 8; 9. median is (5+6)/2 = 5.5 (the number is in ascending order
already)
b. 1; 3; 5; 6; 9. Median is 5
c. 89; 90; 45; 34; 33; 10. Median is (45 + 34)/2 = 39.5

d. 89; 90; 45; 34; 39; 10; 100. (they are not in the ascending order)

Sort data: 10; 34; 39; 45; 89; 90; 100. Median is 45.
Ascending order: theo thứ tự từ bé đến lớn.

e. 89; 90; 45; 30; 33; 10.

Sort data in ascending order: 10; 30; 33; 45; 89; 90.
Median is (33+45)/2

f. 10; 30; 33; 45; 89; 90; 95; 100000.

Median = (45+89)/2
Outlier: 100000

g. 10; 30; 33; 45; 89; 90; 95;100.

Median =(45+89)/2
 Median is not affected by the outlier.

2. Extreme value (outliers)

0,0005; 10; 34; 39; 45; 89; 90; 1900; 25000.
1900 is extreme value because this is much different from the rest.
Mean = (89 + 90 + 45 + 34 + 39 + 10 + 100 + 1900)/8
Median =

3. Mode
a. 1; 2; 3; 4; 5.
No mode
b. 1; 2; 3; 4; 4; 5; 6; 7; 7; 8; 8; 9.
Mode is 4; 7; 8.

4. Midrange
Midrange = ( xmax + xmin)/2
 Midrange is affected by outlier.
CHAPTER 5

5.40 The following contingency table shows average yield (rows) and average duration
(columns) for
38 bond funds. For a randomly chosen bond fund, find the probability that: BondFunds
a. The bond fund is long duration.
b. The bond fund has high yield.
c. The bond fund has high yield given that it is of short duration.
d. The bond fund is of short duration given that it has high yield.

a. P(D3) = 15/38
b. P(Y3) = 15/38
c. P(Y3/D1) = P (Y3 and D1) / P(D1) = 2/38 : 11/38 = 2/11
d. P(D1/Y3) = P(D1 and y3) / P(Y3) = 2/38 : 15/38 = 2/15

*5.99 Dolon Web Security Consultants requires all job applicants to submit to a test for illegal
drugs. If the applicant has used illegal drugs, the test has a 90 percent chance of a positive
result. If the applicant has not used illegal drugs, the test has an 85 percent chance of a
negative result. Actually, 4 percent of the job applicants have used illegal drugs.
If an applicant has a positive test, what is the probability that he or she has actually used illegal
drugs? Hint: Make a 2 * 2 contingency table of frequencies, assuming 500 job applicants.

Summary:
FALSE NEGATIVE :
Those who use the illegal drug but the outcome is negative.
Those who get pregnant but the test is negative.

FALSE POSITIVE
Those who do not use illegal drug but the outcome is positive.
Those who do not get pregnant but the test is positive.

Illegal drug: 90% P

Not use illegal drug: 85% N
4% illegal

Positive (P) Negative (N)

Use illegal (U) = 90% of 4% = 0.004 4%
= 0.9*0.04
= 0.036
Not use illegal (N) = 0.96 – 0.816 =85% of 96% =100% - 4%
= 0.144 = 0.85 * 0.96 = 0.96
= 0.816
= 0.18 = 0.82 100%

P ( U/P) = P(U and P) / P(P) = 0.036 / 0.18 = 0.2

(Condition đứng liền sau if, given that, as long as…)

Based on the tree diagram, we know:

P (get flu/ received flu vaccine) = 0.1
P (don’t get flu/ received flu vaccine) = 0.9
P (get flu/ no received flu vaccine) = 0.4
P (don’t get flu/ no received flu vaccine) = 0.6

QUESTION?
P(get flu and get flu vaccine ) = 0.7 * 0.1 = 0.07

5.41
Of grocery shoppers who have a shopping cart, 70 percent pay by credit/debit card (event C1),
20 percent pay cash (event C2), and 10 percent pay by check (event C3).

Of shoppers without a grocery cart, 50 percent pay by credit/debit card (event C1), 40 percent
pay cash (event C2), and 10 percent pay by check (event C3).

On Saturday morning, 80 percent of the shoppers take a shopping cart (event S1) and 20
percent do not (event S2).
(a) Sketch a tree based on these data.
(b) Calculate the probability of all joint probabilities (e.g., S1 and C1).
(c) Verify that the joint probabilities sum to 1.

SUMMARY:

With grocery cart:

70% credit/debit card (C1)
20% cash (C2)
10% check (C3)

Without grocery cart

50% credit/debit card (C1)
40% cash (C2)
10% check (C3)

Shopping cart = 80% (S1)

Not shopping cart = 20% (S2)
 CHAPTER 6

1. Công thức chỉ cho P(X ≤ x)

Vậy P(X ≥ x) làm như thế nào?

S = a …….. 28, 29, 30, 31, ….. b

P (S) = P(a) + …. + P(28) + P(29) + P(30) + P(31) + … P (b) = 1

Ví dụ:
P(X ≥ 30) = P (X = 30) + P(X = 31) + P(X=32)+…= 1 – P(X ≤ 29) = 1 – P(X< 30)
P(X > 30 ) = 1 – P (X ≤ 30)
P(X < 31) = P ( X ≤ 30)
P (X ≤ 31) (KHÔNG CẦN CHUYỂN ĐỔI)

Sample space: {….. 27, 28, 29,30, 31. , ……. 56, 57…}

2. Binomial Distribution
- Eg: Students in class: 35. Among them, 5 fail and 30 pass
Question: what is probability that students pass?
P (success)? = 30/35

- Eg: Students in class: 35. Among them, 5 fail and 30 pass

Question: what is probability that students fail?
P (success)? = 5/35

6.23 In the Ardmore Hotel, 20 percent of the customers pay by American Express
credit card.
(a) Of the next 10 customers, what is the probability that none pay by American
Express?
(b) At least two?
(c) Fewer than three?
(d) What is the expected number who pay by American Express?
(e) Find the standard deviation.
(f) Construct the probability distribution (using Excel or Appendix A).
(g) Make a graph of its PDF, and describe its shape.

SUMMARY:
2 groups:
20% pay by American Express credit card
80% don’t pay by American Express credit card
n = 10
pi = 20%
 (a) P (X = 0) (KHÔNG NÊN SỬ DỤNG HÀM SUM)
(b) P (X >=2) (NÊN SỬ DỤNG HÀM SUM)
(c) P (X<3) (NÊN SỬ DỤNG HÀM SUM)
(d) Mean = n*pi (expected number)
(e) SD?
EXPLANATION
(a)

(b) P(X>=2) (X từ 2 đến 10) = 0.6241

CASIO 570 shift log

(c) P (X<3) (X từ 0 đến 2) = 0.677

(d) Mean = n* pi = 10*20% = 2
(e) SD = (n*pi*(1 – pi))^(1/2) = √ 10∗0.2∗0.8 = 1.2649

6.47 ABC Warehouse has eight refrigerators in stock. Two are side-by-side models and six are
topfreezer models.
(a) Using Excel, calculate the entire hypergeometric probability distribution for the number
of top-freezer models in a sample of four refrigerators chosen at random.
(b) Make an Excel graph of the PDF for this probability distribution and describe its
appearance.
SUMMARY:
2 outcomes = side-by-side model and topfreezer models (the refrigerators have side by side and
topfreezer models)
N=8
n=4
s=6

4 refrigerators (there are 3 situations)

a. No of Top freezer = 0
No of side by side = 4
(wrong because side by side = 2)

b. No of side by side = 0
No of top freezer = 4

c. No of side by side = 1
No of top freezer = 3

d. No of side by side = 2
No of top freezer = 2

P (X = 2) + P(X=3) + P(X=4)
 CHAPTER 7
1. Phân biệt PDF và CDF
P (X = a): PDF
P (X < a):
P (X > a):
P (X <= a): CDF
P (X >= a):

2. Phân biệt ch6 và ch7

Ch6: “<” khác “<=”
Ch7: giống nhau

3. U (3,6). Find mean and SD?

UNIFORM CONTINUOUS DISTRIBUTION
Lower limit: 3 (a)
Upper limit: 6 (b)

7.15 Find the standard normal area for each of the following, showing your reasoning clearly
and indicating which table you used.
a. P (-1.22 < Z < 2.15)
- C1: P (-1.22 < Z < 2.15) = P(-1.22 < Z < 0) + P(0 < Z < 2.15) = 0.3888 + 0.4842 =
0.873
- C2: P (-1.22 < Z < 2.15) = P ( Z < 2.15) – P (z < -1.22) = 0.9842 – 0.1112 = 0.873

b. P(-3.00 < Z < 2.00)

c. P(Z < 2.00)
d. P(Z = 0)

Ex: P(-1.84 < Z < 2.44)

C2: P(-1.84 < Z < 2.44) = 0.9927 – 0.0329
Ex: P (-2.3 < Z < 3.27) = 0.99946 – 0.0107

Ex: P ( z < -0.67) (su dung bang C1) =

7.53 In Santa Theresa, false alarms are received at the downtown fire station at a mean
rate of 0.3 per day.
(a) What is the probability that more than 7 days will pass before the next false alarm
arrives?
(b) Less than 2 days?
(c) Explain fully

Summary:
False alarms: events
Day: time
0.3 false alarms/day: mean events per unit of time (lamda)
probability that more than 7 days will pass before the next false alarm arrives. ->
EXPONENTIAL

Giả sử hỏi what is probability that in a selected random day,

(a) no false alarm arrives? -> POISSON
(b) more than 3 false alarms arrive? -> POISSON

Eg: during a day at my hospital, there are average 5 clients visiting.

What is the probability no client come today?

Events: clients
Time: a day
Lamda: 5 clients/day
What is the probability no client come today?
 POISSON

What is the probability more than 8 days before the next client visiting?
  EXPONENTIAL
CHAPTER 8

1. Confidence interval for mean with unknown SD

Khi được cho 1 sample bao gồm các số, không cho gì thêm thì phải tự tính sample mean
và sample standard deviation (dùng casio và excel)

2. Confidence interval và Confidence width có phải là một không?

- Confidence interval # Confidence width

3. Phân biệt Confidence interval with a known SD and with the unknown SD
a. The fuel economy of a 2011 Lexus RX 350 2WD 6 cylinder 3.5 L automatic 5-speed using
premium fuel is normally distributed with a known standard deviation of 1.25 MPG. If a
random sample of 10 tanks of gas yields a mean of 21.0 MPG, find the 95 percent
confidence interval for the true mean MPG. (Source: www.fueleconomy.gov.)
1.25 MPG: population SD

b. Guest ages at a Vail Resorts ski mountain typically have a right-skewed distribution.
Assume the standard deviation (σ) of age is 14.5 years. (a) Even though the population
distribution of age is rightskewed, what will be the shape of the distribution of X−, the
average age, in a random sample of 40 guests? (b) From a random sample of 40 guests,
the sample mean is 36.4 years. Calculate a 99 percent confidence interval for μ, the true
mean age of Vail Resorts ski mountain guests.

14,5 years: population SD

c. A sample of 21 minivan electrical warranty repairs for “loose, not attached” wires (one
of several electrical failure categories the dealership mechanic can select) showed a
mean repair cost of $45.66 with a standard deviation of $27.79. (a) Construct a 95
percent confidence interval for the true mean repair cost. (b) How could the confidence
interval be made narrower? (Data are from a project by MBA student Tim Polulak.)

$27.79: Sample SD

4.

a. Mean
 b. Standard deviation

5. Tìm critical value của t có 2 cách

a. Dùng bảng D
 b. Dùng excel.

Eg: nếu đề cho 95% confidence level, và sample size bằng 10

Thì dùng T.INV.2T (probability, df)
Trong đó probability là alpha. Vì 1-alpha = 95% nên alpha bằng 5%
Còn df = n-1 = 10-1 = 9
 CHAPTER 9
1. P value method
Right tailed: P value = P (z > Z calc)
Left tailed: P value = P (z < z calc) (z calc is negative)
Two-tailed:
p value = 2P (z < z calc) (z calc is negative)
OR p value = 2P (z > z calc) (z calc is positive)

(lớn hơn số dương, nhỏ hơn số âm)

NOTE: tính test statistics: nghĩa là tính z calc hoặc là t calc.

9.25 The lifespan of xenon metal halide arc-discharge bulbs for aircraft landing lights is
normally distributed with a mean of 3,000 hours and a standard deviation of 500 hours.
If a new ballast system shows a mean life of 3,515 hours in a test on a sample of 10
prototype new bulbs, would you conclude that the new lamp’s mean life exceeds the
current mean life at α=.01? What is the p-value? (For more information, see
www.xevision.com.)

Summary:
Hypothesized mean = 3000 hours
Populayion SD = 500 hours
Sample mean = 3,515
Sample size = 10
Alpha = 0.01

Hypothesis:
H0: Sample mean ≤ Hypothesized mean
H1: Sample mean > Hypothesized mean

Hypothesis:
H0: Sample mean ≤ 3,000
H1: Sample mean > 3,000 (right-tailed)

NOTE:
- Muốn biết là right-tailed, left-tailed, hay là two-tailed, thì phải đọc đề để xem có
dấu gì. Nếu có chứa dầu “=” thì nó ở H0, còn lại thì nằm ở H1.
- khi đọc đề, nếu thấy more than hoặc less than, thì dấu > hoặc < này nằm ở H 1
- khi đọc đề, nếu thấy at least (eg: 4 mins or more) hoặc at most (eg: 4 mins or
less), thì dấu ≥ hoặc ≤ này nằm ở H0.

9.24 The Scottsdale fire department aims to respond to fire calls in 4 minutes or less, on
average.
Response times are normally distributed with a standard deviation of 1 minute. Would a sample
of 18 fire calls with a mean response time of 4 minutes 30 seconds provide sufficient evidence
to show that the goal is not being met at α = .01? What is the p-value? (See Arizona Republic,
November 23, 2006, p. A10.)

Summary
Hypothesized mean = 4
Population SD = 1
Sample size = 18
Sample mean = 4 mins and 30 secs

Hypothesis:
H0: sample mean ≤ 4 mins
H1: sample mean > 4 mins (right-tailed)
 CHAPTER 12
1. EXERCISE
In a sample of size n = 36, a sample correlation of r = -.350 provides sufficient evidence to
conclude that the population correlation coefficient differs significantly from zero in a two-
tailed test at:
both α = .01 and α = .05.
neither α = .01 nor α = .05.
α = .01
α = .05

tcalc = r([(n - 2)/(1 – r^2)]^(1/2)) = (.-350)([(36 - 2)/(1 – (-0.35)^2)]^(1/2)) = -2.1786 < t.05 = -
2.032 for d.f. = 36 - 2 = 34. However, the test would not be significant for t.01 = - 2.728

2. William used a sample of 78 large U.S. cities to estimate the relationship between Crime
(annual property crimes per 100,000 persons) and Income (median annual income per
capita, in dollars). His estimated regression equation was Crime = 530 + 0.045 Income. If
Income decreases by 800, we would expect that Crime will:

decrease by 36.

remain unchanged.

increase by 450.

decrease by 50