Study Notes on Confidence Intervals: Small Samples and Proportions

Key Concepts in Confidence Intervals

Introduction to Confidence Intervals

• Discussion of estimating confidence intervals with small sample sizes.

• Distinction between small sample sizes (less than 30 cases) and large sample sizes (30 or more cases).

Small Sample Sizes

• When sample size < 30:

  • Considered a small sample, cannot use the standard normal (z) distribution.

  • Reasons: Assumptions of normal distribution do not apply to small samples due to the Central Limit Theorem.

• Use of t-distribution:

  • Definition: A probability distribution used for estimating population parameters when the sample size is small.

  • Characteristics:

  • Bell-shaped and symmetrical.

  • Centers around a mean of zero.

  • Flatter and has heavier tails than the standard normal distribution.

  • Specific to sample size, defined by degrees of freedom.

  • If sample size increases (e.g., n >= 120), the t-distribution approaches the z-distribution.

Degrees of Freedom (df)

• Calculation: $df = n - 1$ where n is the sample size.

• Important for using the t-table when constructing confidence intervals.

Constructing Confidence Intervals with t-distribution

• Process involves constructing intervals using the formula:

  • For small sample sizes, instead of finding z-scores, find t-values based on df.

• Example scenarios include:

  • A 95% confidence level corresponds to an alpha ($\alpha$) level of 0.05.

  • Queries involving confidence intervals always considered two-tailed tests.

• Reference to using t-table:

  • Locate degrees of freedom in the left column and the corresponding alpha level on the top row to find the critical t-value.

Example of Small Sample Calculation

• Sample of 10 students with 95% confidence interval:

  • Degrees of freedom: $df = 10 - 1 = 9$.

  • Critical t-value = 2.262 (found from t-table).

• Sample of 25 students with a 99% confidence level:

  • Degrees of freedom: $df = 25 - 1 = 24$.

  • Corresponding critical t-value = 2.797.

Formula for Confidence Intervals

• General formula for confidence intervals with means:

  • $ext{Confidence Interval} = ext{mean} \, ext{±} \, (t_{critical} imes ext{Standard Error})$

• Calculation of Standard Error (SE):

  • $ext{SE} = rac{s}{ ext{sqrt}(n)}$ where s is the sample standard deviation.

Intuitive Understanding of Confidence Intervals

• The confidence interval estimates population parameters from sample data, showing the range of values likely to contain the true population parameter.

• Interpretation requires understanding significance of the interval produced:

  • e.g., “We can be 95% confident that the mean work overload is between 29 and 37.”

Confidence Intervals for Proportions

• Transition from confidence intervals for means to proportions:

  • Formula adjustment occurs for proportions, requiring conditions for sample size confirmation:

  1. n imes p > 5 where p is the sample proportion.

  2. n imes (1 - p) > 5 so that both conditions must be satisfied to use z-distribution.

• Example survey with a community concerning police dissatisfaction:

  • If 33% of 100 surveyed reported dissatisfaction, then need respective proportions and calculations to construct the confidence interval.

Practical Application in Research

• Importance of demonstrating work through all steps and showing calculation rigor during work with confidence intervals.

• Evaluate and interpret results critically, noting the precision of the interval and implications for generalizations made about broader populations.

• Emphasis on the clarity of communication regarding results, especially within polling or research contexts.