Section 7.2 Confidence Intervals for the Mean When σ Is Known

Confidence Intervals - Section 7.2

Introduction to Confidence Intervals

  • Explains the basics of confidence intervals, especially when the population standard deviation (σ) is known.

  • Covers concepts, calculations, and practical applications of constructing confidence intervals for the mean.

Assumptions for Finding a Confidence Interval

  • Two assumptions for computing a confidence interval for the mean when σ is known:

    1. The sample must be a random sample.

    2. Either the sample size n ≥ 30 or the population is normally distributed when n < 30.

Robust Statistical Techniques

  • Some techniques are considered 'robust' and can tolerate slight deviations from normality.

  • Even with non-perfect normal distribution, confidence intervals can still yield valid conclusions.

Formula for Confidence Interval of Mean (σ Known)

  • General formula for computing confidence intervals for the mean:

    • 90% confidence interval: 𝑧𝛼/2 = 1.65

    • 95% confidence interval: 𝑧𝛼/2 = 1.96

    • 99% confidence interval: 𝑧𝛼/2 = 2.58

Example 7-1: Skipping Breakfast

  • Researcher estimates number of times a person skips breakfast annually:

    • Sample: 49 people

    • Mean: 50 times

    • σ = 2.6

    • 95% confidence interval calculation: (49.3, 50.7)

Example 7-2: Amount of Candy Consumed

  • Nutritionist finds average amount of candy consumed annually:

    • Sample size: 60 adults

    • Mean: 23.9 pounds

    • σ = 4.8

    • 99% confidence interval calculation: (22.3, 25.5)

Diagram for 95% Confidence Interval

  • Figure 7-3 explains how 95% of the sample means fall within the designated range.

Introduction to Confidence Intervals (continued)

  • Confidence intervals provide a range for estimating population parameters.

  • Knowing population standard deviation (σ) allows for higher accuracy in calculations.

Confidence Levels and Critical Values

  • Confidence levels (90%, 95%, 99%) affect the critical values (z-scores) used in calculations.

  • Critical value 𝑧𝛼/2 represents the cutoff point for desired confidence levels.

Finding z(α/2) for a 98% Confidence Interval

  1. Set α = 1 - 0.98 = 0.02.

  2. Divide α by 2: α/2 = 0.01 (z-score corresponds to 98% of data in normal distribution).

Using the Standard Normal Table

  1. Locate area closest to 0.9900 (for 98% confidence level).

  2. Find corresponding z-score, approximately 2.33.

When to Use Positive z Scores in Formulas

  • Only positive z-scores are used in confidence interval formulas due to the symmetry of the normal distribution around the mean.

Normal Distribution and Sample Size

  • When σ is known and sample size (n) is large (n ≥ 30), standard normal distribution applies.

  • Central Limit Theorem states distribution of sample means approaches normality as n increases.

Determining Sample Size

  • Use the formula to determine sample size needed for a specific margin of error (E), ensuring confidence interval meets the desired confidence level.

EXAMPLE 7-3: Credit Union Assets

  • Data represents random sample of assets (in millions) from 30 credit unions:

    • Assume population standard deviation is 14.405.

    • Task: Find the 90% confidence interval of the mean.

    • Sample data: [12.23, 16.56, 4.39, ...] (30 values total).

SOLUTION Step 1: Mean Calculation

  • Calculate mean for the data:

    • Mean X = 11.091

    • Population standard deviation assumed: 14.405.

SOLUTION Step 2: Finding α/2

  • For 90% confidence interval: a = 1 - 0.90 = 0.10.

SOLUTION Step 3: Finding Za/2

  • Z score from normal table for area between 0.9495 and 0.9505 is 1.65 (can also use 1.645 for precision).

SOLUTION Step 4: Substitute in the Formula

  • Confidence range from:[ 11.091 < 11.091 + 4.339 < 15.430 ]

  • Interpretation: 90% confident that population mean assets are between $6.752 million and $15.430 million.

Example 7-4: Automobile Thefts

  • A sociologist estimates average number of automobile thefts per day within a city:

    • Desired confidence: 99%

    • Standard deviation: 4.2

    • Margin of Error: 2

    • Sample size calculation result suggests to survey 30 days to achieve desired confidence.