🎯 Interpreting Confidence Intervals

What "95% Confident" Really Means

CORRECT interpretation: "If we repeated the sampling procedure many times, approximately 95% of the confidence intervals constructed would capture the true population parameter."

In other words: It's about the procedure, not the parameter. The confidence level describes the long-run success rate of the method.

WRONG interpretation (very common): ❌ "There is a 95% probability the parameter is in this interval" ❌ "95% of the data fall in this interval" ❌ "The parameter is definitely in this interval"

Once a CI is calculated, the parameter is either in it or it isn't—the probability is either 1 or 0.

Procedure Interpretation

Think of CI construction like a net:

• Each sample produces a different CI (different $\bar{x}$ or $\hat{p}$)
• 95% of these nets will capture the fish (parameter)
• 5% will miss

We never know if our one net caught the fish, but we know the method works 95% of the time.

Common Misinterpretations to Avoid

1. Reversed confidence: "The population is 95% confident the sample mean is in the interval"

   • ❌ Backwards. We're confident about the population, not that the sample fits.

2. Parameter varies: "There's a 95% chance the parameter is between 45 and 55"

   • ❌ The parameter is fixed (though unknown). The interval varies across samples.

3. Confusing with confidence level: Confidence level (95%) ≠ data range

   • ❌ Don't say "95% of observations fall in this interval"

Overlapping Confidence Intervals vs Hypothesis Tests

When two 95% CIs overlap:

• At the 5% significance level, the difference may not be statistically significant
• But you CAN'T conclude no difference; overlap doesn't guarantee non-significance

Rule of thumb (approximate):

• Non-overlapping 95% CIs → significant at 5% level (two-sided)
• Overlapping CIs → difference is not necessarily non-significant

When to use:

• Overlapping CIs suggest possible non-significance, but do a formal test to be sure
• Non-overlapping CIs strongly suggest significance

Example: Interpreting a Confidence Interval

Survey result: A 95% CI for the proportion of adults who support a policy is (0.52, 0.60).

CORRECT statement: "We are 95% confident that the true proportion of adults supporting the policy is between 52% and 60%. This means if we repeated the survey many times, about 95% of the intervals we construct would contain the true parameter."

INCORRECT statement: "There is a 95% probability that the true proportion is between 52% and 60%." (Once calculated, the parameter is either there or not—no probability involved)

Interpreting Width and Margin of Error

Wider CI:

• Less precise (greater margin of error)
• But higher confidence (if comparing two intervals with same data)
• Results from larger $t^*$ or $z^*$ value OR smaller sample size

Narrower CI:

• More precise (smaller margin of error)
• But lower confidence (if comparing intervals from same data)
• Results from larger sample size or lower confidence level

Factors Affecting CI Width

$\text{Width} = 2 \cdot z^* \cdot SE$

1. Sample size: Larger n → smaller SE → narrower CI
2. Confidence level: Higher confidence → larger $z^*$ → wider CI
3. Population variability: Larger σ → larger SE → wider CI

AP Exam Tip

Interpretation questions require careful language. Never say probability about a fixed parameter; use "confident" or "procedure" language. If asked to compare CIs, discuss precision vs confidence. If asked what would narrow a CI, state: "increase sample size" or "decrease confidence level."