Confidence Intervals
Learn how to estimate a population parameter using a range of plausible values instead of relying on a single point estimate.
Why Do We Need Confidence Intervals?
In the previous lesson, we learned about point estimation. A point estimate gives us one number, such as a sample mean of ₹2,500.
But a sample is only one possible sample from the population. Because of sampling variation, the sample mean may not be exactly equal to the population mean.
A confidence interval gives us a range of values that is constructed from sample data and is intended to capture the unknown population parameter at a stated confidence level.
Point Estimate + Margin of Error = Confidence Interval
What You Will Learn
What a confidence interval means
Why point estimates alone are not enough
Confidence level
Margin of error
Lower and upper confidence limits
The basic confidence interval structure
Correct interpretation of a confidence interval
Real-world analytics applications
From One Number to a Range
Suppose a sample of customers has an average monthly spending of ₹2,500.
Point Estimate
₹2,500
Instead of saying that the population mean is exactly ₹2,500, we can calculate a range such as:
₹2,350 — ₹2,650
Example confidence interval
This range communicates uncertainty caused by sampling variation.
Anatomy of a Confidence Interval
A confidence interval has three important ideas: the point estimate, the margin of error, and the resulting interval.
Example
₹2,500 ± ₹150
Point Estimate
₹2,500
Margin of Error
₹150
Interval
₹2,350–₹2,650
Confidence Interval Formula
At a basic level, a confidence interval can be represented as:
Point Estimate ± Margin of Error
General structure:
Estimate ± Critical Value × Standard Error
The exact critical value depends on the confidence level and the statistical method being used. Later lessons will examine specific confidence intervals in more detail.
What Does 95% Confidence Mean?
A 95% confidence interval is associated with a procedure that, over many repeated random samples under the same conditions, would produce intervals containing the true population parameter about 95% of the time.
Think in terms of repeated samples
Imagine repeatedly taking random samples from the same population and constructing a 95% confidence interval from each sample.
Approximately 95% of intervals would contain the true parameter under the assumptions of the procedure.
Confidence Level
Common confidence levels include 90%, 95%, and 99%.
90%
Lower confidence level generally produces a narrower interval, when other factors are held constant.
95%
A commonly used confidence level in statistical analysis.
99%
Higher confidence generally requires a wider interval, when other factors are held constant.
Important:
Higher confidence is not automatically better. It usually comes with a wider interval, so analysts must consider the purpose of the analysis.
Margin of Error
The margin of error tells us how far the confidence interval extends from the point estimate on each side.
Lower Limit
₹2,350
Point Estimate
₹2,500
Upper Limit
₹2,650
Margin of Error = Upper Limit − Point Estimate
₹2,650 − ₹2,500 = ₹150
What Determines the Width of an Interval?
The width of a confidence interval depends on several factors. Two especially important factors are sample size and variability.
Larger Sample
A larger sample generally reduces standard error and can produce a narrower interval when other factors remain the same.
Greater Variability
Greater variability generally increases uncertainty and can produce a wider interval when other factors remain the same.
How Should You Interpret a Confidence Interval?
Example:
Suppose a 95% confidence interval for the population mean is ₹2,350 to ₹2,650.
A correct frequentist interpretation refers to the long-run performance of the confidence interval procedure: if we repeatedly sampled under the same conditions and constructed intervals in the same way, about 95% of those intervals would contain the true population mean.
Common Mistake
In frequentist statistics, we should not interpret a 95% confidence interval as saying there is a 95% probability that the fixed population parameter is inside this particular interval.
E-Commerce Average Order Value
An e-commerce company selects a random sample of customers to estimate average order value.
Sample Mean
₹2,500
Margin of Error
₹150
Confidence Level
95%
95% CI: ₹2,350 to ₹2,650
The interval gives decision-makers a range rather than relying only on the sample mean of ₹2,500.
Test Your Understanding
CHECK YOUR UNDERSTANDING
What does a confidence interval provide?
CHECK YOUR UNDERSTANDING
A point estimate is ₹2,500 and the margin of error is ₹150. What is the confidence interval?
CHECK YOUR UNDERSTANDING
Which confidence level is commonly used in statistical analysis?
CHECK YOUR UNDERSTANDING
If the confidence level is increased while other factors remain the same, what generally happens to the interval?
CHECK YOUR UNDERSTANDING
What generally happens to the standard error when sample size increases?
CHECK YOUR UNDERSTANDING
What is the correct frequentist interpretation of a 95% confidence interval?
Fill in the Blanks
A confidence interval gives a ______ of plausible values for a population parameter.
Point estimate ± ______ of error gives the basic confidence interval structure.
A commonly used confidence level is ______%.
A larger sample generally produces a smaller standard ______.
Customer Spending Estimate
A sample of customers has a mean monthly spending of ₹3,200. The calculated margin of error is ₹200.
Calculate the confidence interval.
Lower = ₹3,200 − ₹200 = ₹3,000
Upper = ₹3,200 + ₹200 = ₹3,400
Confidence interval = ₹3,000 to ₹3,400
Final Challenge
Two analysts estimate the same population mean.
Analyst A
Estimate = ₹5,000
Margin of error = ₹300
Analyst B
Estimate = ₹5,000
Margin of error = ₹100
Which interval is narrower?
Analyst B's interval is narrower because its margin of error is smaller.
Lesson Summary
A confidence interval estimates a population parameter using a range of values.
A point estimate is the center of the basic interval structure.
The margin of error determines how far the interval extends from the point estimate.
Higher confidence generally produces a wider interval when other factors remain constant.
Larger samples generally reduce standard error and can produce narrower intervals.
A 95% confidence level refers to the long-run coverage of the interval-producing procedure.
Confidence intervals communicate uncertainty caused by sampling variation.