0% found this document useful (0 votes)
3 views5 pages

Sample Answer

The document covers various statistical concepts including conditional expectation, significance testing, t-tests, z-tests, power calculations, p-values, confidence intervals, and F-tests. It provides formulas and examples for calculating expected values, rejection criteria, and power of tests, along with specific numerical cases for clarity. Additionally, it discusses the implications of different rejection cutoffs and the expected number of confidence intervals containing a population mean.
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
3 views5 pages

Sample Answer

The document covers various statistical concepts including conditional expectation, significance testing, t-tests, z-tests, power calculations, p-values, confidence intervals, and F-tests. It provides formulas and examples for calculating expected values, rejection criteria, and power of tests, along with specific numerical cases for clarity. Additionally, it discusses the implications of different rejection cutoffs and the expected number of confidence intervals containing a population mean.
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

1) Conditional expectation from a joint table

Given:
𝑃(𝑋 = 1, 𝑌 = 1) = 0.1, 𝑃(1,2) = 0.1, 𝑃(2,1) = 0.3, 𝑃(2,2) = 0.5.

sampleQuestions-fin_econStat

Pattern

• 𝑃(𝑌 = 𝑦) = ∑! 𝑃(𝑥, 𝑦)
"(!,%)
• 𝑃(𝑋 = 𝑥 ∣ 𝑌 = 𝑦) = "('(%)
• 𝐸(𝑋 ∣ 𝑌 = 𝑦) = ∑! 𝑥 𝑃(𝑋 = 𝑥 ∣ 𝑌 = 𝑦)

(a) 𝑬(𝑿 ∣ 𝒀 = 𝟏)

𝑃(𝑌 = 1) = 0.1 + 0.3 = 0.4


𝑃(𝑋 = 1 ∣ 𝑌 = 1) = 0.1/0.4 = 0.25
𝑃(𝑋 = 2 ∣ 𝑌 = 1) = 0.3/0.4 = 0.75

𝐸(𝑋 ∣ 𝑌 = 1) = 1(0.25) + 2(0.75) = 1.75

(b) 𝑬(𝒀 ∣ 𝑿 = 𝟐)

𝑃(𝑋 = 2) = 0.3 + 0.5 = 0.8


𝑃(𝑌 = 1 ∣ 𝑋 = 2) = 0.3/0.8 = 0.375
𝑃(𝑌 = 2 ∣ 𝑋 = 2) = 0.5/0.8 = 0.625

𝐸(𝑌 ∣ 𝑋 = 2) = 1(0.375) + 2(0.625) = 1.625

2) “More accurate test” when comparing significance/power


Statement (i): if 𝑇) has higher significance level than 𝑇* but same power → “more accurate” =
smaller significance (smaller Type I error)
✅ Answer (i): 𝑇*

sampleQuestions-fin_econStat

Statement (ii): same significance, but 𝑇) has higher power everywhere under 𝐻) → it dominates
✅ Answer (ii): 𝑇)

sampleQuestions-fin_econStat
3) One-mean t-test, decide reject/not reject (two versions)
Given: 𝑛 = 16, 𝑋¯ = 7, 𝑠 = 7, 𝜇+ = 9.

sampleQuestions-fin_econStat

Pattern

𝑋¯ − 𝜇+
𝑡=
𝑠/√𝑛

Here: 𝑠/√𝑛 = 7/4 = 1.75, so

7−9
𝑡= = −1.142857. . .
1.75

At 5% one-sided, critical is about ±1.75(they literally give 𝑃(𝑇 > 1.75) = 0.05).

sampleQuestions-fin_econStat

(i) 𝑯𝟏 : 𝝁 < 𝟗(left-tailed)

Reject if 𝑡 < −1.75.


But −1.143 > −1.75⇒ ✅ Not reject

(ii) 𝑯𝟏 : 𝝁 > 𝟗(right-tailed)

Reject if 𝑡 > 1.75.


But −1.143is not > 1.75⇒ ✅ Not reject

4) One-mean z-test power


Given: 𝑛 = 4, population 𝑁(𝜇, 7* ), test 𝐻+ : 𝜇 = 1vs 𝐻) : 𝜇 > 1, 𝛼 = 0.05. Ask power at 𝜇 = 2.

sampleQuestions-fin_econStat

Pattern
• Standard error: 𝜎/√𝑛 = 7/2 = 3.5
• Reject if

𝑋¯ − 𝜇+
𝑍= > 𝑧+.-. ≈ 1.648
𝜎/√𝑛

So cutoff for 𝑋¯:

𝑋¯ > 1 + 1.648(3.5) = 1 + 5.768 = 6.768

Power at 𝜇 = 2:

6.768 − 2
𝑃/(* (𝑋¯ > 6.768) = 𝑃 ⁣(𝑍 > ) = 𝑃(𝑍 > 1.362) ≈ 0.09
3.5

(Using their normal tail values, it lands around 0.09–0.10.)

✅ Power ≈ 0.09

5) Two tests with different rejection cutoffs


𝑇) : reject iff 𝑅 > 10
𝑇* : reject iff 𝑅 > 18

sampleQuestions-fin_econStat

Pattern

• Bigger cutoff ⇒ reject less often ⇒ smaller significance and smaller power

So:

• 𝛼) > 𝛼*
• 𝛽) > 𝛽* (power at 𝜇 = 1)

✅ 𝛼) > 𝛼* , 𝛽) > 𝛽*
6) p-value decision
Test: 𝐻+ : 𝜇 = 0.04vs 𝐻) : 𝜇 < 0.04. p-value = 0.04.

sampleQuestions-fin_econStat

Pattern
Reject iff 𝑝 ≤ 𝛼.

• At 5%: 0.04 ≤ 0.05⇒ reject


• At 1%: 0.04 ≰ 0.01⇒ not reject

✅ Correct choice: (C)

sampleQuestions-fin_econStat

7) Confidence interval repetition + F-test


Q7-1 Expected number of CIs containing 𝝁𝑨

Given: known 𝜎0 = 8, 𝑛 = 100, CI last day (494.32,496.28), repeated 200 days.

sampleQuestions-fin_econStat

Half-width:

496.28 − 494.32 1.96


𝑚= = = 0.98
2 2

SE = 𝜎/√𝑛 = 8/10 = 0.8

So 𝑧 \* ≈ 𝑚/𝑆𝐸 = 0.98/0.8 = 1.225→ confidence level ≈ 𝑃(∣ 𝑍 ∣≤ 1.225) ≈ 0.78.


Using their provided table values you’d round toward 𝑧 \* ≈ 1.28 ⇒ 80%.

Expected containing:

200 × 0.80 = 160

✅ About 160 intervals (accept 156–160 depending on rounding).


Q7-2 Two-variance (F) test at 5%

Given: 𝑛0 = 15, 𝑠0 = 5; 𝑛1 = 12, 𝑠1 = 8. Test 𝐻+ : 𝜎0 = 𝜎1 vs 𝐻) : 𝜎0 ≠ 𝜎1 .

sampleQuestions-fin_econStat

Pattern

𝑠0* 25
𝐹= = = 0.3906
𝑠1* 64

df: (14,11).

Two-sided 𝛼 = 0.05⇒ 𝛼/2 = 0.025each tail.


They give upper 0.025 critical for 𝐹(14,11): 6.36.

sampleQuestions-fin_econStat

Lower 0.025 critical = 1/𝐹+.+*. (11,14) = 1/6.09 ≈ 0.164(they give 6.09for 𝐹(11,14)).

sampleQuestions-fin_econStat

Reject if 𝐹 < 0.164or 𝐹 > 6.36.


But 0.3906is between ⇒ ✅ Not reject 𝐻+

You might also like