Confidence Intervals and Central Limit Theorem
Key Terms
Population: Collection of subjects with characteristic of interest.
Sample: Subset from the population.
Parameters: Numerical quantities characterizing the population's measurement distribution.
Point Estimates (Statistic): Numerical value from sample estimating a parameter.
Confidence Intervals
Definition: Plausible range of values for the population parameter.
Importance: Confidence intervals improve the accuracy of estimates by providing a range.
Interpretation of Confidence Level: E.g., 95% confidence means approximately 95% of intervals constructed from repeated samples contain the true mean (µ).
Central Limit Theorem (CLT)
Definition: The sampling distribution of the sample mean approaches a normal distribution as sample size (n) increases (n ≥ 30).
Conditions for CLT: Requires random sampling and independence among observations.
Constructing Confidence Intervals
Formula: Point Estimate ± z* × SE, where z* is the z-value for desired confidence level.
Example for Mean: For 95% confidence, use z* = 1.96.
Interpretation: The resulting interval provides a range of values within which the true population parameter is likely to fall.
Proportion Confidence Intervals
Formula: For sample proportion, use
ext{p} ext{±} z* imes ext{SE} ,
where SE is based on sample size.Application Example: If proportion p = 0.80 from a sample of 1000 adults, CI could be calculated as (0.78, 0.82).
Common Misconceptions
Clarification: Saying a parameter falls within the confidence interval does not imply a probability; the interval represents our uncertainty about the estimate, not the probability of the parameter falling within it.
Impact of Confidence Level on Interval Width
Observation: Increasing confidence requires a wider interval for precision.
Drawbacks: Wider intervals may be less informative.
Summary**
Point estimates from random samples give best guesses for parameters.
Confidence intervals reflect the accuracy and variability of these estimates, relevant to the central limit theorem.