ChatGPT Share

QUANTS (1)(2).pdf
PDF

from this pdf itself give me detailed and structured notes for
reading 8 hypotheis testing and make sure to iclude very
possible information and preparing me for the best and
about numericals, formula and how can I calculate it o
calculator as well , tl me everything.

Yes — I can create complete, exam-ready notes for Learning Module 8: Hypothesis
Testing directly from your CFA Quant PDF (LM 8 starts around page 217). Since you want
detailed notes with formulas, numericals, calculator steps, and all important
concepts, I’ll structure this like high-quality study notes you can revise from before exams.

Learning Module 8 — Hypothesis Testing (CFA Quant)

From your PDF: Learning Module 8: Hypothesis Testing QUANTS (1)

1. What is Hypothesis Testing?

Hypothesis testing is a statistical method used to decide whether sample evidence
supports a claim about a population.

In finance, we use it to test things like:

Is portfolio return > benchmark?

Has risk changed?
Is mutual fund manager adding value?
Are two assets equally risky?

Basic idea:
You make a claim → collect sample data → test if evidence supports claim.

Example:
A fund manager says:

“My fund gives more than 12% annual return.”

We test whether this statement is statistically valid.

2. Important Terms (Must Memorize)
Population Parameter
True value for entire population.

Examples:

Population mean return = μ

Population variance = σ²
Population std dev = σ

Sample Statistic
Calculated from sample data.

Examples:

Sample mean = x̄
Sample variance = s²
Sample std dev = s

3. Null and Alternative Hypothesis

This is the heart of hypothesis testing.

Null Hypothesis (H₀)

Default statement.
Usually contains equality.

Examples:

μ = 10%
μ ≤ 10%
σ² = 25

Alternative Hypothesis (H₁ or Hₐ)

What you want to prove.

Examples:
 μ ≠ 10%
μ > 10%
μ < 10%

Rule:
Equality sign always goes in H₀.

Examples:

Test if mean return exceeds 8%

H₀: μ ≤ 8%
H₁: μ > 8%

4. One-tailed vs Two-tailed Test

Right-tailed test
Testing “greater than”

H₁: μ > μ₀

Use when checking increase.

Example:
Return > 12%

Left-tailed test
Testing “less than”

H₁: μ < μ₀

Use when checking decrease.

Example:
Risk < 20%
Two-tailed test
Testing “different”

H₁: μ ≠ μ₀

Use when checking change in either direction.

Example:
Return changed from 10%

Visualization:

Left tail ← rejection region

Middle → accept H₀
Right tail → rejection region

5. Type I and Type II Errors

Very important for theory questions.

Type I Error
Reject true H₀.

Meaning:
You conclude something changed, but actually nothing changed.

Probability = α

Called significance level

Type II Error
Fail to reject false H₀.

Meaning:
Real change happened but you missed it.

Probability = β
Power of Test
Power = 1 − β

Higher power = better test.

Easy memory:
Type I = False Alarm
Type II = Missed Detection

6. Significance Level (α)

Common levels:
10% = 0.10
5% = 0.05
1% = 0.01

Smaller α = stricter evidence required.

Example:
At 5% significance:

There’s only 5% chance of Type I error.

7. Test Statistic
This converts sample data into a number for comparison.

General formula:

Sample Statistic - Hypothesized Value

Test Statistic =
Standard Error
​

Everything in hypothesis testing comes from this.

8. Decision Rules
Two methods:
Method 1: Critical Value Approach
Compare:

Test statistic vs critical value

Example:
At 5% right-tailed Z-test:

Critical value = 1.645

Decision:

If test statistic > 1.645 → Reject H₀

Else → Do not reject H₀

Method 2: p-value Approach

p-value = probability of getting result as extreme as sample.

Decision:

p ≤ α → Reject H₀
p > α → Do not reject H₀

Example:

α = 5%
p = 0.03

Since:

0.03 < 0.05

Reject H₀

9. Z-Test for Mean

Use when:
Population variance known OR
Large sample (n ≥ 30)
Formula:

ˉ − μ0
x
Z=
​

σ/ n ​

Where:
x̄ = sample mean
μ₀ = hypothesized mean
σ = population std dev
n = sample size

Calculator Steps (Scientific Calculator)

Suppose:

x̄ = 52
μ₀ = 50
σ=8
n = 64

Formula:

52 − 50
Z=
8/ 64
​

Step 1:
√64 = 8

Step 2:
8/8 = 1

Step 3:
(52−50)/1 = 2

Answer:

Z=2

Numerical Example 1
A portfolio claims mean annual return = 15%.

Sample:
x̄
 x̄ = 16.2%
σ = 4%
n = 64

Test at 5%.

Step 1 Hypothesis
H₀: μ = 15%
H₁: μ ≠ 15%

Two-tailed

Step 2 Test Statistic

16.2 − 15
Z=
4/ 64
​

1.2
=
4/8
​

1.2
=
0.5
​

= 2.4

Step 3 Critical Value

5% two-tailed:

Critical values:

±1.96

Step 4 Decision
2.4 > 1.96

Reject H₀

Conclusion:
Mean return is significantly different from 15%.
10. t-Test
Use when:
Population variance unknown
Small sample

Formula:

ˉ − μ0
x
t=
​

s/ n ​

Difference from Z:

Use sample std dev (s).

Degrees of Freedom
df = n − 1

Example:
n = 25

df = 24

Used to find t critical value.

Numerical Example 2
Given:

x̄ = 12
μ₀ = 10
s=5
n = 25

12 − 10
t=
5/ 25
​

2
=
5/5
​
 2
=
1
​

=2

t=2

Then compare with t-table.

11. Tests of Return in Finance

Common CFA application:

Testing whether manager beats benchmark.

Example:

Benchmark return = 8%

Hypothesis:

H₀: μ ≤ 8%
H₁: μ > 8%

If test statistic falls in rejection region:

Manager significantly beats benchmark.

12. Tests of Risk (Variance Test)

Sometimes we test variance/risk.

Example:
Did volatility increase?

Use Chi-square test.

Formula:

(n − 1)s2
2
χ =
σ02
​

Where:
s² = sample variance
σ₀² = hypothesized variance
Numerical Example 3
Given:

n = 20
s² = 36
σ₀² = 25

19 × 36
χ2 =
25
​

684
=
25
​

= 27.36

Compare with chi-square critical values.

13. Test for Difference Between Means (Dependent

Samples)
Used when same subject measured twice.

Examples:

Before vs after training

Return before strategy vs after strategy

These are paired observations.

Steps
Compute differences:

di = Xi − Yi
​ ​ ​

Then calculate:
Mean difference = d̄
Std dev of differences = s_d

Formula:

dˉ
t=
sd / n
​

​ ​
Numerical Example 4
Differences:

2, 4, 3, 1, 5

Mean:

2+4+3+1+5
dˉ = =3
5
​

Suppose:

s_d = 1.58

n=5

3
t=
1.58/ 5
​

√5 = 2.236

3
=
0.707
​

= 4.24

Compare with t critical.

14. Equality of Two Variances (F-Test)

Used to compare risks of two assets.

Example:
Is stock A more volatile than stock B?

Formula:

s21
F= 2
​

s2 ​

Put larger variance on top.

So:

F ≥1
Numerical Example 5
Asset A variance = 49
Asset B variance = 25

49
F=
25
​

= 1.96

Compare with F critical.

15. Parametric vs Nonparametric Tests

Parametric Tests
Require assumptions about population distribution.

Examples:

Z-test
t-test
Chi-square
F-test

Assumptions:
Often normal distribution.

Nonparametric Tests
Do not assume distribution.

Use when:

Data not normal

Small sample
Ordinal/rank data

Examples:
Sign test
Rank test
16. When Use Nonparametric Tests?
Use when:

✓ Data highly skewed

✓ Outliers exist
✓ Distribution unknown
✓ Rank data available

Advantages:
Fewer assumptions
Flexible

Disadvantages:
Less powerful than parametric tests

17. Full Hypothesis Testing Process (Exam Steps)

Memorize this sequence.

Step 1
State hypotheses

H₀ and H₁

Step 2
Choose significance level α

Usually 5%

Step 3
Choose test statistic
Z
t
 χ²
F

Step 4
Calculate statistic

Use formula

Step 5
Find critical value / p-value

Step 6
Decision

Reject or do not reject

Step 7
Conclusion in words

Example:

“There is sufficient evidence to conclude…”

18. Critical Values to Memorize

Very important for numericals.

Z critical values

One-tailed
10% → 1.282
5% → 1.645
1% → 2.33

Two-tailed
 10% → ±1.645
5% → ±1.96
1% → ±2.576

19. Calculator Tips for Fast Exam Solving

For any numerical:

Square Root
Use √ button

Example:

36 = 6
​

Standard Error
Formula:
σ
SE = ​

n ​

Example:

σ=12, n=36

Calculator:

√36 = 6
12 ÷ 6 = 2

SE = 2

Z value
Compute numerator first.

Example:

(52 − 50) = 2

Then denominator.
Avoid doing everything together.

20. Common Exam Traps

Trap 1
Using Z instead of t

Remember:

Known σ → Z
Unknown σ → t

Trap 2
Wrong tail

Check wording carefully:

greater than → right tail

less than → left tail
different → two-tail

Trap 3
Wrong variance order in F-test

Always:

Larger variance / smaller variance

21. Ultra-Short Revision Sheet

Memorize this night before exam:

Hypothesis:
H₀ = null
H₁ = alternative
Errors:
Type I = reject true H₀ = α
Type II = fail to reject false H₀ = β

Tests:

Z-test:

ˉ−μ
x
​

σ/ n ​

t-test:

ˉ−μ
x
​

s/ n ​

Chi-square:

(n − 1)s2
σ2
​

F-test:

s21 ​

s22
​

Decision:
test statistic beyond critical value → Reject H₀

How to Score Best in LM 8

Focus in this order:

Highest priority
✅ Hypothesis setup
✅ Type I / II errors
✅ Z & t numericals
✅ Chi-square
✅ F-test
✅ Tail identification
Medium priority
✅ p-value
✅ Nonparametric tests
✅ Interpretation questions

This chapter is very formula-heavy, but once you understand the decision logic,
numericals become easy.

If you want, I can also make a one-shot CFA revision PDF / formula sheet for LM8
for last-minute revision.