Unit 06 - Confidence Intervals

Definition: Statistical inference is the process of using sample data to make general conclusions (inferences) about a population.
Data is often collected from a sample, but there is a desire to infer information about the larger population of interest.
Objective: The goal of statistical inference is to make statements about a population parameter based on observed sample data.

Confidence Intervals: Estimation of a population parameter, specifically the population mean ( $\mu$ ).
Tests of Significance: Assessing evidence for claims about population parameters.

Statistical inference employs the language of probability to quantify uncertainty, acknowledging that a sample may or may not represent the population accurately.
Good sampling methods, such as Simple Random Sampling (SRS), enhance the validity of conclusions made.

An interval is constructed around a sample mean to estimate the population parameter ( $\mu$ ).

A confidence interval is defined as an interval estimate that combines a point estimate with a margin of error to reflect uncertainty in estimation.
Example: If sample mean ( $\bar{x}$ ) is used to estimate the population mean ( $\mu$ ), we aim to create a range such that $\mu$ lies within it with a specified level of confidence.

The sample mean is centered at $\bar{x}$ , and with the 68-95-99.7 rule, approximately 95% of observations from a normal distribution lie within two standard deviations of the mean. - Mathematical Representation:
- If $\bar{x} \sim N(\mu, \sigma/\sqrt{n})$ , then about 95% of available $\bar{x}$ values fall within $2\frac{\sigma}{\sqrt{n}}$ of $\mu$ .
This leads to the formulation of the 95% confidence interval for $\mu$ : - $\bar{x} \pm 2\frac{\sigma}{\sqrt{n}}$

Given: - Standard deviation ( $\sigma$ ) = 0.40
- Sample mean ( $\bar{x}$ ) = 3.31
- Sample size ( $n$ ) = 25
Calculation of 95% confidence interval for true mean GPA: - $\bar{x} \pm 2\frac{0.40}{\sqrt{25}}$
- $3.31 \pm 0.16$
- Interval: (3.15, 3.47)

A 95% confidence interval means that if we took numerous samples and created intervals the same way, about 95% of those intervals would include the true mean ( $\mu$ ).

\text{Estimate \pm Margin of Error} - Estimate: The sample mean ( $\bar{x}$ ).
- Margin of Error: For a 95% CI, it is represented as $2\frac{\sigma}{\sqrt{n}}$ .

Each confidence interval has an associated confidence level (C) that denotes the probability the interval captures the true population mean ( $\mu$ ).
Common confidence levels utilized: 90%, 95%, etc.
We typically aim for a higher confidence level for improved estimation accuracy.

The value of $z^*$ is critical to find confidence intervals.
For example, for 90% confidence, we derive: - Area in each tail = 5% (0.05)
- Total area for central 90% = 0.90
- $P(Z < z^*) = 0.05 + 0.90 = 0.95$
- This leads to critical value $z^* = 1.645$ (from normal distribution tables).

92% Confidence Interval:
- Sample mean = 117.8, critical value $z^* = 1.75$ .
- Resulting CI: $117.8 \pm \frac{1.75 \times \sigma}{\sqrt{10}}$ .
- Final CI interval calculated reflects the range of values.
99% Confidence Interval:
- Take critical value $z^* = 2.576$ directly from tables.
- Calculated interval reflects the length and width with increased certainty but reduced precision in estimation.

Increasing sample size leads to reduced margins of error thereby narrowing the confidence interval.
The relationship states that increasing sample size by a factor of k results in a marginal error reduction by k.
For example, for our earlier example, increasing from 40 to 160 decreases the margin of error significantly, leading to more accurate results.

A sufficient sample size must be determined based on the desired precision and confidence level.
Formula for sample size regarding margin of error ( $m$ ) is used to establish optimal size for studies, indicated as follows: - $n = \left(\frac{z^* \sigma}{m}\right)^2$

Scenario planning for adults' sleep hours, requiring n for different confidence scenarios and margins of error leads to calculations reflecting sample adjustments based on confidence thresholds.

Common misconceptions about confidence intervals include: - Misconceptions that intervals apply to individuals rather than population means.
- Misunderstandings regarding the relation of the sample mean to confidence intervals leading to inaccurate probability statements.
- Misconceptions about the relationship between sample means across repeated samples misinterpreted as applying to fixed values of population means.

Confidence intervals provide an essential statistical tool to estimate population parameters while allowing for quantification of sampling uncertainty. Usage of adequate sample sizes and appropriate margin of errors optimally informs the derivation of such estimates. The final assessments made via confidence intervals rest on foundational principles which require rigorous adherence to proper sampling methodologies to ensure their validity.