Chapter 2: Confidence Interval Estimation
1. Point vs Interval Estimates
Type Definition Example
Example: the sample
A single best guess of a population
Point estimate mean X̄ = 50 estimates
parameter (based on a sample).
μ
Interval estimate A range that likely contains the true “We are 95% confident
(Confidence population parameter with a specified that μ is between 45
Interval) confidence level. and 55.”
2. Confidence Intervals/ interval estimate
Point Estimate±(Critical Value)(Standard Error)\
A confidence interval gives a range of values:
️Formula Components:
Point estimate: sample statistic (mean X̄ or proportion p)
Critical value: depends on confidence level (Z or t value)
Standard Error (SE): variability of the estimate
⚠️Confidence level (1 – α)
Represents the percentage of all possible intervals that contain the true parameter.
For example, a 95% confidence level means that 95% of such intervals will contain
μ.
3. Confidence Interval for the Mean (σ Known)
4. Confidence Interval for the Mean (σ Unknown)
Since σ is rarely known, we use the sample standard deviation (s) and the t-
distribution.
Assumptions
Population standard deviation is unknown
Population is normally distributed If population is not normal, use large
sample
Note:
The t-distribution has fatter tails than Z (to account for extra uncertainty).
As n increases, t → Z.
Confidence t t t
Z (∞)
Level (df=10) (df=20) (df=30)
90% 1.812 1.725 1.697 1.645
95% 2.228 2.086 2.042 1.96
99% 3.169 2.845 2.750 2.58
The tα/2 value depends on degrees of freedom (d.f.)
Number of observations that are free to vary after sample mean has been calculated
5. Confidence Interval for a Proportion (π)
Used when estimating the population proportion (e.g., % of customers satisfied).
6. Determining Sample Size
For the Mean:
The margin of error is also called sampling error
the amount of imprecision in the estimate of the population parameter
the amount added and subtracted to the point estimate to form the confidence
interval
For the Proportion:
If π unknown → use π = 0.5 for maximum variability.
statistical sampling in auditing
1. Time and Cost Efficiency – Sampling is quicker and less expensive than
examining the entire population.
2. Objective Sample Size Determination – Statistical methods allow auditors to
calculate the required sample size in advance objectively.
3. Defensible Results – Because sampling relies on statistical principles, audit
findings are objective, defensible, and can be justified to superiors or in court.
4. Estimation of Sampling Error – Sampling allows auditors to quantify the
potential error, enabling generalization of findings to the entire population with a
known level of accuracy.
5. Improved Accuracy for Large Populations – For large datasets, sampling
often yields more reliable conclusions than reviewing every item, which may
introduce non-sampling errors.
6. Combining Multiple Samples – Auditors can merge and evaluate samples
collected by different individuals, improving the overall assessment of the
population.
10. Ethical Guidelines
When reporting confidence intervals:
Always include sample size, confidence level, and margin of error.
Always interpret what the interval means in words.
Avoid giving only the point estimate without explaining its uncertainty.
11. Ambiguous Terms Explained
Term Meaning
Significance level (probability that the true value lies outside the
α (alpha)
CI). Example: α=0.05 → 5% error chance.
Zα/2 Z value that leaves α/2 in each tail (e.g., 1.96 for 95% CI).
tα/2 t-distribution value depending on degrees of freedom.
df (degrees of
n – 1 → number of values that can vary freely in the sample.
freedom)
Standard error Measure of variability in the sample estimate. Smaller SE →
(SE) narrower CI.
Margin of error
Half the width of the CI → determines precision.
(e)
Confidence level
How sure we are that the interval contains the true value
(1–α)