You are on page 1of 24

Hypothesis Testing and Confidence Intervals

假设检验和置信区间

1-24

FRMFinancial Risk Manager)金融风险管理师

Mean and Variance

  • Population Mean

  • Sample Mean

  • Population Variance

  • Sample Variance

 

N

X

i

i=1

μ=

N

n

 

X

i

 

X=

i=1

n

 

N

 

X

i

 

2

=

i=1

 
 

n

X

i

2

S =

i=1

 

n-1

 

N

μ

2

X

2

2-24

FRMFinancial Risk Manager)金融风险管理师

Central Limit Theorem

For simple random samples of size n from a population with a mean µ and a variance σ², the sampling distribution of the sample mean approaches N(µ, σ²/n) if the sample size is sufficiently large (n 30).

X

N (

,

 2 ) n
2
)
n

Standard error of the sample mean

  • Known population variance

  • Unknown population variance

   / n x s x  s / n
   /
n
x
s
x 
s
/
n

3-24

FRMFinancial Risk Manager)金融风险管理师

Point Estimation and Confidence Interval Estimate

Point Estimation

X

Confidence Interval Estimation

  • Level of significance (α)

  • Degree of Confidence (1α)

  • Confidence Interval = [ Point Estimate ± (reliability factor) × Standard error]

4-24

FRMFinancial Risk Manager)金融风险管理师

Confidence Interval Estimate

  • If the population has a normal distribution with a known variance, a confidence interval for the population mean can be calculated as:

X

z

σ α 2 n
σ
α 2
n
Confidence Interval Estimate  If the population has a normal distribution with a known variance ,

5-24

X

 1.96 n
1.96
n
  • X X

95%

 1.96
 1.96

n
n

FRMFinancial Risk Manager)金融风险管理师

真题回顾

Answer:B The confidence interval for the interest rate coefficient is: [1.50-1.96×0.20,1.50+1.96×0.20]= [1.11, 1.89]
Answer:B
The confidence interval for the interest rate coefficient is:
[1.50-1.96×0.20,1.50+1.96×0.20]= [1.11, 1.89]

6-24

FRMFinancial Risk Manager)金融风险管理师

Example

  • The 95% confidence interval of the sample mean of employee age for a major corporation is 19 years to 44 years based on a z-statistic. The population of employees is more than 5,000 and the sample size of this test is 100. Assuming the population is normally distributed, the standard error of mean employee age is closest to:

    • A. 1.96.

    • B. 2.58.

    • C. 6.38.

    • D. 12.50.

Example  The 95% confidence interval of the sample mean of employee age for a major

7-24

FRMFinancial Risk Manager)金融风险管理师

Confidence Interval Estimate

  • If the distribution of the population is normal with unknown variance, we can use the t-distribution to construct a confidence interval:

s X  t α 2 n
s
X
t
α 2
n
Confidence Interval Estimate  If the distribution of the population is normal with unknown variance, we

8-24

X

   2 n
 
2
n
  • X X 2

95%

Confidence Interval Estimate  If the distribution of the population is normal with unknown variance, we

n
n

FRMFinancial Risk Manager)金融风险管理师

Hypothesis Testing

Hypothesis: a statement about the value of a population parameter

to be tested The null hypothesis (H 0 ) and alternative hypothesis (H a ) One-tailed test vs. Two-tailed test

  • One-tailed test

H 0 : μ≥0 H 0 : μ≤0

H a : μ<0 H a : μ>0

  • Two-tailed test H 0 : μ=0

H a : μ≠0

9-24

FRMFinancial Risk Manager)金融风险管理师

Test Statistic and Critical Value

Test statistic test statistic =

sample statistic - hypothesized value

standard error of the sample statistic

Critical value

  • The distribution of test statistic (z, t, χ², F)

  • Significance level (α)

  • One-tailed or two-tailed test

10-24

FRMFinancial Risk Manager)金融风险管理师

Decision Rule

Reject H 0 if |test statistic|>critical value Fail to reject H 0 if |test statistic|<critical value

2.5% 2.5% 95% -1.96 1.96 Reject H 0 Fail to Reject H 0 Reject H 0
2.5%
2.5%
95%
-1.96
1.96
Reject H 0
Fail to
Reject H 0
Reject H 0
5% 95% 1.645 Fail to Reject H 0 Reject H 0
5%
95%
1.645
Fail to
Reject H 0
Reject H 0

We can never say acceptH 0 State the conclusion: μ is (not) significantly different from μ 0

11-24

FRMFinancial Risk Manager)金融风险管理师

P value testing

2.5% 2.5% 1.07% 1.07% Negative of the test statistic Critical value for 5% significance level Positive
2.5%
2.5%
1.07%
1.07%
Negative of the
test statistic
Critical value for 5%
significance level
Positive of the
test statistic
  • P - value=2.14%

If P–value < alpha, we reject null hypothesis
If P–value < alpha, we reject null hypothesis

12-24

FRMFinancial Risk Manager)金融风险管理师

Test of Single Population Mean

  • H 0 : μ=μ 0

  • z-test vs. t-test

 

Normal population, n<30

n≥30

Known population variance (σ²)

z-test

z-test

Unknown population variance

t-test

t-test or z-test

  • z-statistic

  • t-statistic

x   Z  0  / n x   t  0 n
x
 
Z 
0
/
n
x  
t
0
n
1
s /
n

13-24

FRMFinancial Risk Manager)金融风险管理师

Test of Single Population Variance

  • H 0 : σ²0 ²

  • The chi-square test (χ²-test)

2

(n

1)s

2

2

0

d

f =n-1

Test of Single Population Variance  H : σ ² =σ ²  The chi-square test

n=sample size s 2 =sample variance σ 0 2 =hypothesized value for the population variance

14-24

FRMFinancial Risk Manager)金融风险管理师

Test of Variances Difference

  • H 0 : σ 1 ²2 ²

  • The F-test

F

s

2

1

s

2

2

df 1 =n 1 -1; df 2 =n 2 -1

  • Always put the larger variance in the numerator (

s

2

1

>

s

2

2

)

  • The rejection region is always the right-side tail, no matter the test is one-tailed or two-tailed

15-24

FRMFinancial Risk Manager)金融风险管理师

Summary of Hypothesis Testing

Test type

Assumptions

H 0

Test-statistic

 

Critical value

 
 

Normally distributed population,

known population variance

μ=0

Z

x

0

 

N(0,1)

 
 / n
/
n
 

Mean

       

hypothesis

Normally distributed population,

μ=0

   

t(n-1)

 

testing

 

x

0

unknown population variance

 

t

s / n
s
/
n
   
 

Normally distributed population

σ²0 ²

 

2

(

n

1)

s

2

 

2

Variance

2

 

(

n

1)

hypothesis

 

0

 
       

testing

Two independent normally distributed populations

σ 1 ²2 ²

 

F

s 2 1 s 2 2
s
2
1
s
2
2

F

(

n

1

1,

n

2

1)

16-24

FRMFinancial Risk Manager)金融风险管理师

真题回顾

17-24

FRMFinancial Risk Manager)金融风险管理师

真题回顾

  • 72. An analyst believes that the average return of hedge funds was12% over the past year. The analyst picks a random sample of 10 hedge funds and finds that their average return was 29% with a standard of 24% over the past year. The analyst decides to test at a 95% confidence level the null hypothesis H 0 : the average return of hedge funds was 12%. The possible related critical values of the t-statistic are provided below:

真题回顾 72. An analyst believes that the average return of hedge funds was12% over the past

The conclusion drawn from the t-test should be to:

  • A. Reject the null hypothesis since the t-statistic is greater than the critical value.

  • B. Reject the null hypothesis since the t-statistic is less than the critical value.

  • C. Fail to reject the null hypothesis since the t-statistic is greater than the critical value.

  • D. Fail to reject the null hypothesis since the t-statistic is less than the critical value.

18-24

FRMFinancial Risk Manager)金融风险管理师

真题解答

x   29%  12% H 0 : μ=12% H a : μ ≠ 12%
x 
29%
12%
H 0 : μ=12%
H a : μ ≠ 12%
t
0
2.2399
2.262
n  1
s
/
n
24% /
10

19-24

FRMFinancial Risk Manager)金融风险管理师

Type I and Type II Errors

  • Type I error: reject the null hypothesis when its actually true

  • Type II error: fail to reject the null hypothesis when its actually false

  • Significance level (α): the probability of making a Type I error Significance level =P (Type I error)

  • Power of a test: the probability of correctly rejecting the null hypothesis when it is false Power of a test = 1−P(Type II error)

Type I and Type II Errors  Type I error : reject the null hypothesis when

20-24

FRMFinancial Risk Manager)金融风险管理师

真题回顾

真题回顾 FRM ( Financial Risk Manager )金融风险管理师
真题回顾 FRM ( Financial Risk Manager )金融风险管理师

21-24

FRMFinancial Risk Manager)金融风险管理师

Chebyshev’s Inequality

  • 对任何一组观测值,个体落于均值周围k个标准差之内的概率 不小于1 - 1/k²,对任意k>1

  • P(|X-µ|kσ) 1 - 1/k²

k>1

At

least

1

2

2

1

4

3

4

1

1

75%

1

1

3

2

1

8

    89%

1

9

9

1

1

4

2

1

 

1

16

15

16

94%

Lie

within

2

3

4

Standard

deviations

of the mean

22-24

FRMFinancial Risk Manager)金融风险管理师

Example

For a skewed distribution, what is the minimum percentage of the observations that will lie between ±2.5 standard deviations of the mean based on Chebyshev's inequality?

  • A. 56%

  • B. 75%

  • C. 84%

  • D. 95%

Example  For a skewed distribution, what is the minimum percentage of the observations that will

23-24

FRMFinancial Risk Manager)金融风险管理师

结 束

恭 祝 大 家

FRM学习愉快!

顺利通过考试!

24-24

FRMFinancial Risk Manager)金融风险管理师