Exam Question 1

If length is normally distributed with mean 175 and standard deviation 20, what
is the probability that a person is less than 160? Round your answer to two
decimals.

Answer: 0.23

The z score for 160 is: (160-175)/20 = -15/20 = -0.75.

We seek P( z < -0.75). This is equal to P( z > 0.75). We use the table to find ( z >
0.75) = 0.2266. Thus, P( z < -0.75) = 0.2266. The answer is 0.23.

Exam Question 2

In a box there are five red socks, seven blue socks, and eight green socks. If you
randomly select three socks (without replacement), what is the probability you
get red, blue, and green in that order? Round your answer to two decimal places.
Write your answer as a fraction not as a percentage (i.e. write as 0.38 not as
38%).

Answer: 0.04.

The probability of red is initially 5/(5+7+8) = 5/(20) = ¼.

The probability of blue next is 7/19. The probability of green next is 8/18.
Multiply these together and round off, one gets 0.04.

Exam Question 3

If the probability that a professor correctly writes a slide is 0.80, what is the
probability that at least 6 out of 8 slides are correct? Round your answer to two
decimal places. Write your answer as a fraction not as a percentage (i.e. write as
0.38 not as 38%).

Answer: 0.80.

To answer this using the table, we focus on the number of incorrect slides. The
probability that a slide is incorrect is 0.2. We draw a graph to see what
probability we seek:

correct 0 1 2 3 4 5 6 7 8
incorrect. 8 7 6 5 4 3 2 1 0

We seek Pr(at least 6 correct). This is equal to 1-Pr(at least 3 incorrect).

Pr(at least 3 incorrect) can be found using the table, focusing on at least 3 out of
8 and probability 0.2, which gives Pr(at least 3 incorrect) = 0.2031. Thus, Pr(at
least 6 correct) = 1-Pr(at least 3 incorrect) = 0.7969.
Exam Question 4

To check how many hours students spend on Instagram, a sample of 81 students

were taken. The mean number of hours spent on Instagram per day was 2.3 and
the standard deviation was 2.4. Create an approximate 95% confidence interval
around the mean. What is the upper limit of the confidence interval? Round your
answer to one decimal.

Answer: 2.8

The upper limit of the confidence interval is mean + 2*STEM, where STEM =
standardDeviation/sqrt(n).

We get: 2.3+2*(2.4/sqrt(81)) = 2.3+2*(2.4/9) = 2.3+2*0.2666 = 2.8333.

If an exact critical value of 1.96 is used, we get 2.3+1.96*0.2666= 2.82266

Exam Question 5
A company believes that 60% of customers prefers design A over design B. A
sample of 120 customers is taken and 85 prefer design A. Test the null
hypothesis that 60% prefer design A. What is the absolute value of the test
statistic (the “t-value” or “t-statistic”)? Round your answer to one decimal.

Answer: 2.4

Here STEP = sqrt(0.6*0.4 / 120) = sqrt ( 0.24 / 120) = 0.04472.

The observed proportion is 85/120 = 0.70833.
The absolute value of test statistic then becomes: (0.70833-0.6)/ 0.04472 =
2.4234.

Question 6.

There are 35 students in class A. On average their exam grade was equal to 62
with a standard deviation equal to. 12.4. There are 44 students in class B. On
average their exam grade was equal to 57.3 with a standard deviation equal to
10.4. Test the hypothesis that the average grade of the two classes is the same,
using a significance level of 5%. What is the absolute value of the test-statistic
(“t”)? Round your answer to one decimal.

Answer: 1.8

𝑛) 𝑠)+ + 𝑛+ 𝑠++ 1 1 35 × 12.4+ + 44 × 10.4+ 1 1

𝑆𝑇𝐸𝐷𝑀 = ' ' + = ' ' +
𝑛) + 𝑛+ 𝑛) 𝑛+ 35 + 44 35 44
= 2.5661

Calculate test statistic:

t = (m1 – m2)/STEDM = (62-57.3)/2.5661 = 1.8316.

Question 7.
One of the numbers 1,2,3,...., 100 is selected randomly (each of the 100 numbers
equally likely). Suppose you know that the selected number is above 66. What is
the probability that it is above 85, given that it is above 66? Round your answer
to two decimal places. Write your answer as a fraction not as a percentage (i.e.
write as 0.38 not as 38%).

Answer: 0.44

The probability that it is above 66 is 34 / 100. The probability that the number of
above 85 is 15/ 100. The probability that the number is above 85 given that it is
above 66 is (15/ 100) / (34 / 100) = 15 / 34 = 0.4412.
Alternatively, there are 34 numbers above 66. 15 of them are above 85, thus
15/34.

Question 8.
Using the data below, test the hypothesis that there is no difference between the
two samples using a Mann-Whitney test and a significance level of 0.05. What is
the value of the test statistic (the lowest u value)? Your answer should be an
integer.

Sample 1 67 13 6 55 43 44 24
Sample 2 15 22 19 42 33 27 52 8

Answer: 21.

Answer: computations can be done as follows

No obs = 19 n1 = 7 n2 = 8

Sample 1 67 13 6 55 43 44 24
Rank 15 3 1 14 11 12 7
R1 = 63 U1 = 35

Sample 2 15 22 19 42 33 27 52 8
Rank 4 6 5 10 9 8 13 2
R2 = 57 U2 = 21

Lowest U is: 21

Data for exam questions 9 and 10

The following data shows the yearly growth rates for four firms. Based on this
data, test the null hypothesis that the average growth rate is equal for all firms,
using an Anova and a significance level of 0.05.

Firm 1 Firm 2 Firm 3 Firm 4

7 8 10 12
3 5 12 11

Exam Question 9
What is the critical value that the F-ratio should be compared with (round to 2
decimal places)?

Answer: 6.59

Exam Question 10
What is the F-ratio (use at least 2 decimals in each calculation and round the
answer to 1 decimal places)?

Answer: 5.6
 Answer: Computations can be done as follows n= 8
Group 1 Group 2 Group 3 Group 4
obs obsˆ2 obs obsˆ2 obs obsˆ2 obs obsˆ2
7 49 8 64 10 100 12 144
3 9 5 25 12 144 11 121
Sums 10 58 13 89 22 244 23 265

Sum of squares (Sˆ2) = 656 Total (T)= 68 No. Groups (k) = 4

Based on the table we have:

SS_Between = 63, SS_Tot =78, SS-Within = 15

MS_Between = 63/(4-1) = 21
MS_Within = 15/(8-4) = 3.75

F = 21/ 3.75= 5.6

Critical value (F value for signif. 0.05, df_between = 3, df_within = 4) is 6.59
Because F <Critical Value, the null hypothesis is Not rejected.
Data for exam question 11

The following data shows the yearly rate of return for three traders. Based on this
data, test the null hypothesis that the average rate of return is equal for all traders,
using an Anova and a significance level of 0.05.

Trader 1 Trader 2 Trader 3

11 5 9
6 5 8
12 3 7

Exam Question 11
What is the F-ratio (use at least 3 decimals in each calculation and round the
answer to 1 decimal place)?

Answer: 5.3

Answer: Computations can be done as follows n= 9

Group 1 Group 2 Group 3
obs obsˆ2 obs obsˆ2 obs obsˆ2
11 121 5 25 9 81
6 36 5 25 8 64
12 144 3 9 7 49
Sums 29 301 13 59 24 194

Sum of squares (Sˆ2) = 554 Total (T)= 66 No. Groups (k) = 3

Based on the table we have:

SS_Between = 44.667, SS_Tot = 70, SS-Within = 25.333

MS_Between = 44.667/(3-1) = 22.333

MS_Within = 25.333/(9-3) = 4.222

F = 22.333/ 4.222 = 5.289

Critical value (F value for signif. 0.05, df_between = 2, df_within = 6) is 5.14.
Because F >Critical Value, the null hypothesis is rejected.
Data for Exam Questions 12-14

The following data shows performance of four different types of firms marketing
campaigns as a function of the age of the firm (old or young) and whether it was
selling online or just in stores.

Age

old young
stores 9 11
Selling 2 12
format 4 9
Online 7 1
4 4
3 6

Exam Question 12
Using a two factor Anova, what is the F-ratio for the main effect of age? Use at
least 3 decimals in all computations. Round your answer to 1 decimal place.

Answer: 2.5

Exam Question 13
Using a two factor Anova, what is the F-ratio for the main effect of the selling
format? Use at least 3 decimals in all computations. Round your answer to 1
decimal place.

Answer: 6.2

Exam Question 14
Using a two factor Anova, what is the F-ratio for the interaction effect? Use at
least 3 decimals in all computations. Round your answer to 1 decimal place.

Answer: 5.1

In what follows, “a” means “age” and “b” means “selling format”
 Answer: computations can be done like this:
No obs (n)= 12
Levels of factor a (Ka) = 2 Levels of factor b (Kb) = 2
Factor a
a1 a2
obs obsˆ2 obs obsˆ2
9 81 11 121
b1 2 4 12 144
4 16 9 81

Factor b Sums 15 101 32 346 Sum row = 47

obs obsˆ2 obs obsˆ2

7 49 1 1
b2 4 16 4 16
3 9 6 36

Sums 14 74 11 53 Sum row = 25

Sum Column 29 Sum Column 43

Based on the above sum, we can compute the F ratios as follows:

Total (T) = 72
Sum squared (Sˆ2) = 574
SS_a = 140.17 + 308.17 - 432.00 = 16.33
SS_b = 368.17 + 104.17 - 432.00 = 40.33

SS_a*b = 75.00 + 341.33 + 65.33 + 40.33 - 432.00 - 16.33 - 40.33 = 33.33

SS_Tot = 574.00 - 432.00 = 142.00

SS_With = 52.00

MS_a = 16.33 / 1 = 16.33

MS_b = 40.33 / 1 = 40.33
Ms_a*b = 33.33 / 1 = 33.33
Ms_With = 52.00 / 8 = 6.50

F_a = 16.33 / 6.50 = 2.51 Crit. Val is F for 1 and 8 = 5.32

F_b = 40.33 / 6.50 = 6.21 Crit. Val is F for 1 and 8 = 5.32
F_a*b = 33.33 / 6.50 = 5.13 Crit. Val is F for 1 and 8 = 5.32

Exam Questions 15-18

The following table shows grades (y) and past experience (x) for five students.

Student x y
1 3 5
2 9 6
3 2 4
4 1 3
5 7 8

Exam Question 15
What is the correlation coefficient (r) based on this data (round to 1 decimal
place). Use at least 3 decimals in all your computations.

Answer: 0.8

Exam Question 16
Assuming the regression line is Y = a + bX, what is the value of b (round to 1
decimal place)? Use at least 3 decimals in all your computations.

Answer: 0.5

Exam Question 17
Assuming the regression line is Y = a + bX, what is the value of a (round to 1
decimal places)? Use at least 3 decimals in all your computations.

Answer: 3.2

Exam Question 18
Test the null hypothesis that the slope in the regression line is equal to zero,
using a significance level of 0.05 (two-sided test). What is the value of the test-
statistic (the ‘w-value’ or sometimes called ‘t-statistic’)? Round your answer to
one decimal. Use at least 3 decimals in all your computations.

Answer: 2.5

The regression line can be computed as follows

Number of obs. is 5 Signif level: 0.05

obs X X*X Y X*Y Y*Y

1 3 9 5 15 25
2 9 81 6 54 36
3 2 4 4 8 16
4 1 1 3 3 9
5 7 49 8 56 64
Sum 22.00 144.00 26.00 136.00 150.00
Mean 4.40 5.20

SS_xy equals 136.00 - 114.4 = 21.6

SS_x equals 144.00 - 96.8 = 47.2
SS_y equals 150.00 - 135.2 = 14.80
Thus, the correlation coefficient is = 0.82

Thus, b = SS_xy / SS_x = 0.4576271

and a = yhat - b*xhat = 5.20 - 0.4576271 * 4.40 = 3.186441
(yhat is the mean of y and xhat is the mean of x)

Thus the regression line is Ŷ = 3.18644 + 0.458 * X

To compute the test statistic we compute the predicted Y (Ŷ):
Y X Ŷ = a + b * X e = Y-Ŷ (Y-Ŷ)*(Y-Ŷ)
5 3 4.559322 = 3.1864407 + 0.4576271 * 3 0.441 0.194
6 9 7.305085 = 3.1864407 + 0.4576271 * 9 -1.305 1.703
4 2 4.101695 = 3.1864407 + 0.4576271 * 2 -0.102 0.010
3 1 3.644068 = 3.1864407 + 0.4576271 * 1 -0.644 0.415
8 7 6.389831 = 3.1864407 + 0.4576271 * 7 1.610 2.593
Σ (Y-Ŷ)2 4.915254
Thus, Rsq = 1 - ( Σ (Y-Ŷ)2)/SS_y ) = 1 - 4.9152542 / 14.80 = 0.667888

To test the null-hypothesis of zero slope, we compute the test statistic: w = (b-0)/s_b
where s_b*s_b = (Σ (Y-Ŷ)2 / (N-2)) / SS_x

Σ (Y-Ŷ)2 / (N-2) = 4.915254 / 3 = 1.6384181

and SS_x (computed above) is: 47.2

Thus, s_b*s_b = 1.638418 / 47.2 = 0.0347122

and s_b is the squareroot of 0.035 which is 0.1863122

Thus, the test statistic w equals 0.4576271 / 0.1863122 = 2.456

Exam Question 19

The following table shows sales (y) and marketing costs (x) for four firms.

Firm x y
1 3 6
2 7 8
3 2 4
4 5 7

Test the null hypothesis that the slope in the regression line is equal to zero
using a significance level of 0.05 (two-tailed test). What is the value of the test-
statistic (the ‘w-value’ or sometimes called ‘t-statistic’)? Round your answer to
one decimal. Use at least 3 decimals in all your computations.

Answer: 4.1

The regression line can be computed as follows

Number of obs. is 4 Signif level: 0.05

obs X X*X Y X*Y Y*Y

1 3 9 6 18 36
2 7 49 8 56 64
3 2 4 4 8 16
4 5 25 7 35 49

Sum 17.00 87.00 25.00 117.00 165.00

Mean 4.25 6.25

SS_xy equals 117.00 - 106.25 = 10.75

SS_x equals 87.00 - 72.25 = 14.75
SS_y equals 165.00 - 156.25 = 8.75
Thus, the correlation coefficient is = 0.95

Thus, b = SS_xy / SS_x = 0.728814

and a = yhat - b*xhat = 6.25 - 0.7288 * 4.25 = 3.152542
(yhat is the mean of y and xhat is the mean of x)

Thus the regression line is Ŷ = 3.15254 + 0.7288 * X

To compute the test statistic we compute the predicted Y (Ŷ):
Y X Ŷ = a + b * X e = Y-Ŷ (Y-Ŷ)*(Y-Ŷ)
6 3 5.33898 = 3.152542 + 0.7288 * 3 0.661 0.437
8 7 8.25424 = 3.152542 + 0.7288 * 7 -0.254 0.065
4 2 4.61017 = 3.152542 + 0.7288 * 2 -0.610 0.372
7 5 6.79661 = 3.152542 + 0.7288 * 5 0.203 0.041

2
Σ (Y-Ŷ) 0.915254
Thus, Rsq = 1 - ( Σ (Y-Ŷ)2)/SS_y ) = 1 - 0.9153 / 8.75 = 0.8954

To test the null-hypothesis of a slope equal to zero we compute the test statistic: w = (b-0)/s_b
where s_b*s_b = (Σ (Y-Ŷ)2 / (N-2)) / SS_x

Σ (Y-Ŷ)2 / (N-2) = 0.91525 / 2 = 0.4576

and SS_x (computed above) is: 14.75

Thus, s_b*s_b = 0.45763 / 14.75 = 0.031

and s_b is the squareroot of 0.031 which is 0.1761

Thus, the test statistic w equals 0.728814 / 0.1761 = 4.1377

Critical values are 4.30 and -4.30
Thus, the null hypothesis is: not rejected

To compute the test statistic we compute the predicted Y (Ŷ):

Y X Ŷ = a + b * X e = Y-Ŷ (Y-Ŷ)*(Y-Ŷ)
6 3 5.33898 = 3.152542 + 0.7288 * 3 0.661 0.437
8 7 8.25424 = 3.152542 + 0.7288 * 7 -0.254 0.065
4 2 4.61017 = 3.152542 + 0.7288 * 2 -0.610 0.372
7 5 6.79661 = 3.152542 + 0.7288 * 5 0.203 0.041

2
Σ (Y-Ŷ) 0.915254
Thus, Rsq = 1 - ( Σ (Y-Ŷ)2)/SS_y ) = 1 - 0.9153 / 8.75 = 0.8954

To test the null-hypothesis of a slope equal to zero we compute the test statistic: w = (b-0)/s_b
where s_b*s_b = (Σ (Y-Ŷ)2 / (N-2)) / SS_x

Σ (Y-Ŷ)2 / (N-2) = 0.91525 / 2 = 0.4576

and SS_x (computed above) is: 14.75

Thus, s_b*s_b = 0.45763 / 14.75 = 0.031

and s_b is the squareroot of 0.031 which is 0.1761

Thus, the test statistic w equals 0.728814 / 0.1761 = 4.1377

Critical values are 4.30 and -4.30
Thus, the null hypothesis is: not rejected

Exam Question 20
The following data shows how the number of different type of students that have
received distinction or not.
 Grade

distinction pass fail

Course economics 32 21 56
math 29 17 32
accounting 34 46 18

Test the null-hypothesis that there is no association between course and grade
using a Chi-square test. What is the test statistic (the chi-square sum, the
uncorrected one, not the Yates corrected chi-square)? Round your answer to 1
decimal.

Answer: 32.0

Computations can be done like this:

2) Totals: a1 a2 a3 3) Calculate E: a1 a2 a3
b1 32 21 56 109 b1 36.333 32.126 40.540
b2 29 17 32 78 b2 26.000 22.989 29.011
b3 34 46 18 98 b3 32.667 28.884 36.449
95 84 106 285

4a) O-E: a1 a2 a3 4b) (O-E)*(O-E)/E a1 a2 a3

b1 -4.333 -11.126 15.460 b1 0.517 3.853 5.895
b2 3.000 -5.989 2.989 b2 0.346 1.560 0.308
b3 1.333 17.116 -18.449 b3 0.054 10.142 9.338

5) Sum of (O-E)*(O-E)/E = 32.015

No. Columns 3 No. Rows 3
Degree of Freedom = 4
6) Critical Value 9.49
7) Thus, null hypothesis of no association is rejected