More on Normal Distributions

Advanced topics on normal distributions: sampling distribution, central limit theorem, probability calculations, and real-world applications with 5 detailed exercises.

Exercise Solutions 1-3
1 Sampling Distribution of Sample Means
Exercise 1
The heights of adult males are normally distributed with mean 70 inches and standard deviation 3 inches. If we take samples of size 25, what is the mean and standard deviation of the sampling distribution of sample means?
Key Definitions:

Sampling Distribution: The distribution of a statistic (like sample mean) from repeated samples of the same size

Sample Mean: Average of values in a sample

Standard Error: Standard deviation of the sampling distribution

Sampling Distribution Method:
  1. Identify the population parameters (μ, σ)
  2. Apply the sampling distribution formulas
  3. Calculate the mean and standard error of sample means
Step 1: Identify population parameters

Population mean (μ) = 70 inches

Population standard deviation (σ) = 3 inches

Sample size (n) = 25

Step 2: Calculate mean of sampling distribution

Mean of sample means (μₓ̄) = Population mean (μ) = 70 inches

Step 3: Calculate standard error

Standard error (σₓ̄) = σ ÷ √n = 3 ÷ √25 = 3 ÷ 5 = 0.6 inches

Mean of sampling distribution = 70 inches
Standard error = 0.6 inches
Sample size = 25
Final Answer:

The sampling distribution of sample means has a mean of 70 inches and a standard deviation (standard error) of 0.6 inches. The mean remains the same as the population mean, but the standard deviation decreases by a factor of √n, making sample means less variable than individual observations.

Applied Rules:

• Sampling distribution mean: μₓ̄ = μ

• Standard error: σₓ̄ = σ ÷ √n

• Decreased variability: Sample means are less spread out than individual values

2 Central Limit Theorem Application
Exercise 2
The time it takes to complete a standardized test is normally distributed with mean 120 minutes and standard deviation 15 minutes. What is the probability that a randomly selected student completes the test in less than 110 minutes? What about the probability that the mean time for a sample of 36 students is less than 110 minutes?
Key Definitions:

Central Limit Theorem: As sample size increases, the sampling distribution of sample means approaches normality

Individual vs. Sample Mean: Different standard deviations affect probabilities

CLT Application Method:
  1. Calculate probability for individual observation
  2. Calculate probability for sample mean using sampling distribution
  3. Compare the results
Step 1: Individual observation probability

For individual: z = (110 - 120) ÷ 15 = -10 ÷ 15 = -0.67

P(X < 110) ≈ 25% (from standard normal table)

Step 2: Sample mean probability

For sample mean: σₓ̄ = 15 ÷ √36 = 15 ÷ 6 = 2.5

z = (110 - 120) ÷ 2.5 = -10 ÷ 2.5 = -4.0

Step 3: Calculate probabilities

P(Individual < 110) ≈ 25%

P(Sample mean < 110) ≈ 0% (z = -4.0 is extremely unlikely)

Individual: P(X < 110) ≈ 25%
Sample mean: P(X̄ < 110) ≈ 0%
Sample size = 36
Final Answer:

For an individual student, there's approximately a 25% chance of completing the test in less than 110 minutes. However, for the mean time of 36 students, the probability is essentially 0%. This demonstrates the power of the Central Limit Theorem - sample means are much less variable than individual observations.

Applied Rules:

• CLT: Sample means have reduced variability

• Standard error: σₓ̄ = σ ÷ √n

• Reduced probability: Extreme values are less likely for sample means

3 Confidence Interval Construction
Exercise 3
The weights of packages of ground beef are normally distributed with mean 1.0 lb and standard deviation 0.1 lb. What percentage of samples of size 16 will have sample means between 0.95 and 1.05 lb?
Key Definitions:

Confidence Interval: Range of values likely to contain the true parameter

Sampling Variability: How sample statistics vary from sample to sample

Confidence Interval Method:
  1. Calculate the standard error of the sample mean
  2. Convert interval endpoints to z-scores for the sampling distribution
  3. Find the probability between these z-scores
Step 1: Calculate standard error

σₓ̄ = σ ÷ √n = 0.1 ÷ √16 = 0.1 ÷ 4 = 0.025 lb

Step 2: Convert to z-scores

For 0.95: z₁ = (0.95 - 1.0) ÷ 0.025 = -0.05 ÷ 0.025 = -2.0

For 1.05: z₂ = (1.05 - 1.0) ÷ 0.025 = 0.05 ÷ 0.025 = 2.0

Step 3: Calculate probability

P(-2.0 < Z < 2.0) = P(Z < 2.0) - P(Z < -2.0)

= 0.9772 - 0.0228 = 0.9544 = 95.44%

Standard error = 0.025 lb
Z-scores: -2.0 to 2.0
Probability = 95.44%
Final Answer:

Approximately 95.44% of samples of size 16 will have sample means between 0.95 and 1.05 lb. This interval corresponds to 2 standard errors on either side of the population mean, which captures approximately 95% of sample means according to the empirical rule.

Applied Rules:

• Standard error: σₓ̄ = σ ÷ √n

• Sampling distribution: X̄ ~ N(μ, σₓ̄)

• Empirical rule: 95% within 2 standard errors

Exercise Solutions 4-5
4 Normal Approximation to Binomial
Exercise 4
A fair coin is flipped 100 times. What is the probability of getting between 45 and 55 heads? Use the normal approximation to the binomial distribution.
Key Definitions:

Binomial Distribution: Distribution of number of successes in n independent trials

Normal Approximation: Using normal distribution to approximate binomial when n is large

Continuity Correction: Adjusting discrete values for continuous approximation

Normal Approximation Method:
  1. Verify conditions for normal approximation (np ≥ 5, n(1-p) ≥ 5)
  2. Calculate mean and standard deviation of binomial distribution
  3. Apply continuity correction
  4. Standardize and calculate probability
Step 1: Verify conditions

n = 100, p = 0.5

np = 100 × 0.5 = 50 ≥ 5 ✓

n(1-p) = 100 × 0.5 = 50 ≥ 5 ✓

Step 2: Calculate binomial parameters

Mean: μ = np = 100 × 0.5 = 50

Standard deviation: σ = √[np(1-p)] = √[100 × 0.5 × 0.5] = √25 = 5

Step 3: Apply continuity correction and calculate

P(45 ≤ X ≤ 55) becomes P(44.5 < X < 55.5)

z₁ = (44.5 - 50) ÷ 5 = -1.1

z₂ = (55.5 - 50) ÷ 5 = 1.1

P(-1.1 < Z < 1.1) ≈ 72.87%

Mean = 50, SD = 5
Z-scores: -1.1 to 1.1
Probability ≈ 72.87%
Final Answer:

The probability of getting between 45 and 55 heads in 100 flips is approximately 72.87%. The normal approximation is valid here since both np and n(1-p) are greater than 5, and the continuity correction helps bridge the gap between discrete binomial and continuous normal distributions.

Applied Rules:

• Conditions: np ≥ 5 and n(1-p) ≥ 5

• Binomial parameters: μ = np, σ = √[np(1-p)]

• Continuity correction: Adjust discrete bounds by 0.5

5 Percentile and Critical Values
Exercise 5
The lifetime of light bulbs is normally distributed with mean 800 hours and standard deviation 50 hours. What lifetime value separates the top 10% of bulbs from the rest? What is the 90th percentile?
Key Definitions:

Percentile: Value below which a given percentage of observations fall

90th Percentile: Value below which 90% of data falls

Critical Value: Z-score corresponding to a specific probability

Percentile Calculation Method:
  1. Identify the desired percentile (90th percentile)
  2. Find the corresponding z-score from standard normal table
  3. Convert z-score back to original units using x = μ + zσ
Step 1: Identify the percentile

90th percentile means 90% of values fall below this point

Step 2: Find the z-score

From standard normal table, z₀.₉₀ ≈ 1.28

Step 3: Convert to original units

x = μ + zσ = 800 + (1.28)(50) = 800 + 64 = 864 hours

Z-score = 1.28
90th percentile = 864 hours
Top 10% cutoff = 864 hours
Final Answer:

The 90th percentile corresponds to a lifetime of 864 hours. This means that 90% of light bulbs last less than 864 hours, and only 10% last longer than this value. The top 10% of bulbs have lifetimes exceeding 864 hours.

Applied Rules:

• Percentile identification: 90th percentile = top 10%

• Z-score lookup: Use standard normal table

• Conversion formula: x = μ + zσ

Comprehensive Summary: More on Normal Distributions
X̄ ~ N(μ, σ/√n)
Sampling Distribution
Mean
μ
Std Error
σ/√n
Shape
Normal
CLT
n≥30
Key Formulas:

Sampling Distribution: X̄ ~ N(μ, σ/√n)

Standard Error: SE = σ/√n

Normal Approximation: μ = np, σ = √[np(1-p)]

Key Concepts:

Sampling Distribution: The distribution of a statistic (like sample mean) computed from all possible samples of the same size from a population. The sampling distribution of the sample mean is normally distributed if the population is normal or if the sample size is large.

Central Limit Theorem (CLT): States that the sampling distribution of the sample mean approaches a normal distribution as the sample size increases, regardless of the population distribution, provided n ≥ 30.

Standard Error: The standard deviation of a sampling distribution, measuring the variability of the statistic across samples. SE = σ/√n for sample means.

Core Methods and Procedures:
  1. Sampling Distribution Analysis: Calculate μₓ̄ = μ and σₓ̄ = σ/√n
  2. CLT Applications: Use when working with sample means and n ≥ 30
  3. Normal Approximation: Apply to binomial when np ≥ 5 and n(1-p) ≥ 5
  4. Percentile Calculations: Find z-scores and convert back to original units
Tip 1: Remember that sample means have smaller standard deviations than individual observations (divided by √n).
Tip 2: The CLT allows us to use normal distribution even when the population isn't normal, provided the sample size is large enough.
Tip 3: Always check conditions before using normal approximation to binomial: np ≥ 5 and n(1-p) ≥ 5.
Tip 4: Continuity correction is essential when approximating discrete distributions with continuous ones.
Common errors: Confusing population parameters with sampling distribution parameters, forgetting continuity correction, misapplying CLT to small samples.
Memorization aids: "Sampling means are less variable" (√n effect), "CLT kicks in at 30", "Check np and n(1-p) for approximations".
Essential Rules:

• Sampling distribution: μₓ̄ = μ, σₓ̄ = σ/√n

• CLT condition: n ≥ 30 for most distributions

• Normal approximation: np ≥ 5 and n(1-p) ≥ 5

• Continuity correction: Adjust discrete values by ±0.5

Detailed Course Content: More on Normal Distributions
CLT: X̄ ~ N(μ, σ/√n) as n → ∞
Central Limit Theorem
Sampling Mean
μₓ̄ = μ
Sampling distribution mean
Sampling Std Dev
σₓ̄ = σ/√n
Standard error
Normal Approx
μ = np, σ = √[np(1-p)]
Binomial approximation
Detailed Definitions:

Law of Large Numbers: As sample size increases, sample statistics converge to population parameters. The sample mean gets closer to the population mean as n increases.

Finite Population Correction: When sampling without replacement from a finite population, if n/N > 0.05, the standard error is adjusted by multiplying by √[(N-n)/(N-1)].

Consistency: An estimator is consistent if it approaches the true parameter value as the sample size increases.

Advanced Problem-Solving Techniques:
  1. Problem Recognition: Identify whether dealing with individual values, sample means, or proportions
  2. Parameter Selection: Choose correct formulas for sampling distribution parameters
  3. Condition Checking: Verify assumptions before applying approximations
  4. Calculation Strategy: Use appropriate standardization method
  5. Result Interpretation: Contextualize the statistical result
Advanced Tip 1: For proportions, the sampling distribution is approximately normal when np ≥ 5 and n(1-p) ≥ 5, with mean p and standard deviation √[p(1-p)/n].
Advanced Tip 2: The CLT is powerful because it applies to any population distribution, not just normal ones.
Advanced Tip 3: As sample size increases, the sampling distribution becomes narrower and more concentrated around the population mean.
Advanced Tip 4: For small samples from non-normal populations, consider using t-distribution instead of normal distribution.
Real-world applications: Quality control, polling, medical research, finance, engineering, and social sciences rely heavily on these concepts.
Foundation for inference: These concepts are fundamental to confidence intervals, hypothesis testing, and regression analysis.

Questions & Answers

Question: Why does the standard deviation of sample means decrease as sample size increases? It seems counterintuitive.

Answer: This makes sense when you think about it! Sample means tend to cluster around the population mean because extreme values in the sample are balanced by other values. When you take larger samples, these extremes are more likely to cancel out, making the sample mean more stable and closer to the population mean. The standard error formula σₓ̄ = σ/√n shows this mathematically: as n increases, √n increases, so σ/√n decreases. For example, if σ = 10 and n = 4, then σₓ̄ = 5; but if n = 100, then σₓ̄ = 1. Larger samples provide more reliable estimates of the population mean.

Question: What's the difference between the Central Limit Theorem and the Law of Large Numbers?

Answer: While related, they describe different aspects of sampling:

  • Law of Large Numbers: As sample size increases, the sample mean gets closer to the population mean. It's about convergence to a single value.
  • Central Limit Theorem: As sample size increases, the distribution of sample means approaches a normal distribution. It's about the shape of the sampling distribution.

The LLN says the sample mean converges to μ; the CLT says the distribution of sample means becomes normal around μ. Together, they explain why larger samples provide more accurate and reliable estimates.

Question: When should I use the normal approximation to the binomial versus the exact binomial distribution?

Answer: Use the normal approximation when:

  • Both np ≥ 5 and n(1-p) ≥ 5 (these are the conditions for validity)
  • n is large (making exact binomial calculations difficult)
  • You need a quick approximation

Use the exact binomial when:

  • The conditions aren't met (especially when p is very small or very large)
  • You need exact probabilities
  • n is small and exact calculation is feasible

The normal approximation is especially useful for manual calculations when n is large, but with modern computing, exact methods are often preferred for accuracy.