What are Confidence Intervals

Guide: Confidence Intervals

Author's Avatar

Daniel Croft

Daniel Croft is an experienced continuous improvement manager with a Lean Six Sigma Black Belt and a Bachelor's degree in Business Management. With more than ten years of experience applying his skills across various industries, Daniel specializes in optimizing processes and improving efficiency. His approach combines practical experience with a deep understanding of business fundamentals to drive meaningful change.

In statistics, understanding the distinction between ‘statistics’ and ‘parameters’ is crucial. Statistics are the observed values in a sample, such as the measured heights in a group of people. These observations help calculate sample statistics like mean, median, or mode. Parameters, in contrast, are estimated values that represent our beliefs about the general population, like estimating the average height of all people based on a sample.

To infer about a population mean, certain conditions must be met, including random selection of samples, normal distribution, and known standard deviation of the variable. This introduction sets the stage for exploring how statistical methods, particularly confidence intervals, are used to estimate population parameters and understand the uncertainty in these estimates.

Table of Contents

Understanding the Basics

There are two key concepts to understand when dealing with statistics: statistics and parameters. Let’s dissect these.

Statistics (Observed Values): These are the values in our sample that we actually measure or observe. For example, if we were studying a group of people’s heights, the heights we measured would be our observed values or statistics. We frequently use these observed values to calculate things like our sample’s mean (average), median (middle value), or mode (most frequent value).

Parameters (Estimated Values): These are the values that we estimate based on our observed values. They represent our beliefs about the general population. For example, we could estimate the average height of all people based on our sample of people’s heights (not just those in our sample). A parameter is this estimated average height.

Now, let’s talk about the conditions for inference about a mean. Certain conditions must be met before we can make an inference (or educated guess) about a population mean (average):

Random Selection: The sample from which we draw our conclusions must be chosen at random. This means that each individual in the population has an equal chance of being included in the sample. This ensures that our sample is representative of the general population.

Normal Distribution: Ideally, the variable of interest (for example, height) should follow a normal distribution in the population. A normal distribution is a bell-shaped curve with the majority of values clustering around the mean and fewer values at the extremes.

Known Standard Deviation: While the population mean is unknown (that’s what we’re trying to estimate), the standard deviation for the variable should be known. The standard deviation measures how far apart the values are from the mean.

Unfortunately, I am unable to provide graphs in this text-based format; however, examples of normal distribution curves and visual representations of standard deviation can be found online. These visuals can aid in the comprehension of these concepts.

Estimating the Population Mean

We frequently want to use data collected from a sample to make estimates about the larger population. The population mean, or average, is one of the most common things we might want to estimate. This is how we do it:

Confidence Level: The level of confidence indicates how certain we are that our estimate is correct. It is frequently expressed as a percentage. A 95% confidence level, for example, means that we are 95% certain that our estimate is correct. In other words, if we took 100 different samples and calculated 100 different estimates, we would expect approximately 95 of them to be correct.

Confidence Interval: A confidence interval is a range of values that is likely to contain the true population mean. It is calculated using our sample data and the confidence level we have chosen. The confidence interval provides us with a range of possible values rather than a single estimate. This is useful because it informs us about the degree of uncertainty or margin of error in our estimate.

Margin of Error: The margin of error is a measure of the uncertainty in our estimate. It is the amount by which we believe our estimate may be off. The confidence interval is calculated using the margin of error. For example, if our sample mean is 100 and our margin of error is 5, our confidence interval is 95-105 (100 – 5 to 100 + 5). The margin of error is determined by several factors, including sample size and data variability.

Here’s a quick formula for calculating a confidence interval:

    • To begin, compute the sample mean (the average of your sample data).

    • Then, compute your sample’s standard error. This is calculated by dividing the standard deviation by the square root of the sample size.

    • Next, select your level of assurance. The most common options are 90%, 95%, and 99%. Each of these represents a z-score (a measure of how many standard deviations away from the mean you are). A 95% confidence level, for example, corresponds to a z-score of 1.96.

    • To calculate the margin of error, multiply the standard error by the z-score.

    • Finally, compute the confidence interval by adding and subtracting the margin of error from the sample mean.

For example, if your sample mean is 100, your standard error is 2, and you’re using a 95% confidence level (z-score = 1.96), your margin of error would be 1.96 * 2 = 3.92. So your confidence interval would be 100 – 3.92 to 100 + 3.92, or 96.08 to 103.92. This means you can be 95% confident that the true population mean is between 96.08 and 103.92.

The Reasoning of Statistical Estimation

Statistical estimation is a technique for drawing conclusions or making predictions about a population based on data from a sample. This is how it works:

Deriving the Confidence Interval Formula: The formula for a confidence interval is derived from the properties of the normal distribution, which is a bell-shaped curve that describes how data is distributed around the mean. The formula is as follows:

Confidence Interval = Sample Mean ± (Z-Score * Standard Error)

The Z-Score is a number that corresponds to the confidence level you select (for example, 1.96 for a 95% confidence level). The standard error, calculated as the standard deviation divided by the square root of the sample size, is a measure of the variability in your sample data. The margin of error is the amount by which you believe your estimate could be off by the product of the Z-Score and the standard error.

Distribution of Sample Means: If you were to take many samples from the same population and calculate the mean of each sample, those means would form their own distribution, known as the sampling distribution of the mean. The Central Limit Theorem states that if the sample size is large enough, this distribution will be approximately normal, regardless of the shape of the population distribution. This is why we can use the normal distribution’s properties to infer the population mean.

The 68-95-99.7 Rule: This rule, also known as the empirical rule, applies to normal distributions. According to the report, approximately 68% of the data will fall within one standard deviation of the mean, approximately 95% will fall within two standard deviations, and approximately 99.7% will fall within three standard deviations. This rule is used to calculate the Z-Scores for different confidence levels and helps us understand how much data we can expect to fall within certain ranges. A 95% confidence level, for example, corresponds to a range of two standard deviations from the mean, yielding a Z-Score of 1.96.

The 68-95-99.7 rule in the context of confidence intervals helps us understand how confident we can be that the true population mean falls within our interval. A 95% confidence interval, for example, indicates that if we take many samples and calculate an interval for each one, we can expect approximately 95% of those intervals to contain the true population mean.

How Confidence Intervals Behave

Understanding how confidence intervals behave entails understanding the relationship between the confidence level and the margin of error, as well as how to achieve a small margin of error.

Confidence Level and Margin of Error Relationship: The confidence level and the margin of error are inversely related. This means that as the level of confidence rises, so will the margin of error. Why is this the case? A higher level of confidence indicates that you want to be more certain that your interval contains the true population mean. To increase your certainty, widen your interval, which means a larger margin of error. A 99% confidence interval, for example, will be wider than a 95% confidence interval for the same set of data because you want to be more certain of capturing the true mean, so you accept a larger margin of error.

Achieving a Small Margin of Error: The margin of error is influenced by three factors: the population standard deviation, the size of the sample, and the confidence level you select. Here’s how to manage these variables to achieve a lower margin of error:

    • Increase the size of your sample: The larger your sample, the closer your sample mean will be to the population mean, reducing your margin of error. However, keep in mind that the relationship is not linear; for example, doubling the sample size will not cut the margin of error in half. The denominator of the margin of error formula is the square root of the sample size, so you’d need to quadruple your sample size to halve the margin of error.

    • Reduce your confidence level: A lower confidence level results in a smaller margin of error because you’re willing to accept a higher risk that your interval will not contain the population mean. However, this implies that you are less certain of your estimate, which may not be desirable.

    • Choose a population with a lower standard deviation: Although this is out of your hands, populations with less variability (lower standard deviation) will have smaller margins of error because the values are more tightly clustered around the mean.

Remember that achieving a small margin of error frequently necessitates trade-offs. For example, you may need to weigh the precision of your estimate (small margin of error) against your confidence in that estimate (confidence level).

Interpreting Confidence Level

The confidence level is a fundamental concept in statistics that quantifies our level of certainty in a specific statistical conclusion. Here’s how to interpret it:

Understanding the Overall Capture Rate: The overall capture rate is a term that is frequently used to describe the confidence level. This means that if we repeated our sampling process a number of times, generating a confidence interval from each sample, a certain percentage of those intervals would capture (or contain) the true population parameter. For example, a 95% confidence level means that if we took 100 different samples and created a confidence interval from each one, we would expect about 95 of those intervals to contain the true population mean. The other five intervals would not contain the true mean, and this is where our method fails.

What It Means to Have a Certain Level of Confidence: When we say we have a certain level of confidence, we’re expressing how sure we are that our method works. A 95% confidence level, for example, indicates that we are 95% certain that our method of generating a confidence interval will yield an interval containing the true population mean. It is important to note that this does not imply that the true mean is 95% likely to be within our one specific interval. Instead, it means that our method will produce an interval containing the true mean 95% of the time. This is a small but significant distinction.

In summary, the confidence level measures the dependability of our method for generating confidence intervals. A higher confidence level indicates that our method is more reliable, but it also implies that our intervals will be wider (because we are attempting to capture the true mean more frequently), potentially making our estimates less precise.

Practical Application of Confidence Intervals

Using confidence intervals in practice is a methodical process. Here’s a step-by-step procedure:

State the Practical Question: The first step is to clearly define the question you’re trying to answer. This question should be applicable to the data you have. For example, you might want to know the average height of a particular group of people or the average time it takes for a specific chemical reaction to occur.

Plan: Determine the parameter to be estimated (for example, the population mean), select a level of confidence (such as 95%), and select the appropriate type of confidence interval. You’ll also need to collect your data at this point, making sure it’s a random sample and large enough for your needs.

Solve: There are two parts to this step. Check the conditions for the interval you’ve chosen first. This typically entails ensuring that your sample is random, that your sample size is sufficient, and that your data is normally distributed. Second, use the following formula to compute the confidence interval:

Confidence Interval = Sample Mean ± (Z-Score * Standard Error)

The Z-Score corresponds to your chosen confidence level, and the standard error is calculated as the standard deviation divided by the square root of the sample size.

Conclude: Finally, return to your practical question and interpret your results in this context. For example, if your confidence interval for a group’s average height is (160 cm, 170 cm), you could conclude: “We are 95% confident that the average height of individuals in this group is between 160 cm and 170 cm.”


The use of confidence intervals is a cornerstone in statistical analysis, providing a method to estimate population parameters and understand the associated uncertainty. By defining the practical question, planning the analysis, calculating the interval using the sample mean, Z-score, and standard error, and then interpreting the results, we can draw meaningful conclusions about a population.

The relationship between confidence level, margin of error, and sample size plays a critical role in determining the accuracy and reliability of our estimates. Ultimately, the practical application of confidence intervals allows us to make informed decisions and predictions about a larger population based on sample data, striking a balance between precision and certainty in statistical estimations.


A: A confidence interval is a range of values, derived from a data sample, that is likely to contain the value of an unknown population parameter. It provides an estimated range of values which is likely to include an unknown population parameter.

A: A 95% confidence level means that if we were to take 100 different samples and compute a confidence interval for each sample, we would expect about 95 of those intervals to contain the true population mean.

A: The confidence level and margin of error are inversely related. As the confidence level increases, the margin of error also increases. This is because a higher confidence level means a wider interval to ensure that it captures the true population parameter.

A: You can achieve a smaller margin of error by increasing your sample size, decreasing your confidence level, or choosing a population with a smaller standard deviation.

A: A confidence interval is interpreted as a range of values within which the true population parameter lies, with a certain degree of confidence. For example, a confidence interval of (100, 200) at a 95% confidence level means we are 95% confident that the true population parameter is between 100 and 200.


Picture of Daniel Croft

Daniel Croft

Daniel Croft is a seasoned continuous improvement manager with a Black Belt in Lean Six Sigma. With over 10 years of real-world application experience across diverse sectors, Daniel has a passion for optimizing processes and fostering a culture of efficiency. He's not just a practitioner but also an avid learner, constantly seeking to expand his knowledge. Outside of his professional life, Daniel has a keen Investing, statistics and knowledge-sharing, which led him to create the website learnleansigma.com, a platform dedicated to Lean Six Sigma and process improvement insights.

All Posts

Useful Tool

Free Lean Six Sigma Templates

Improve your Lean Six Sigma projects with our free templates. They're designed to make implementation and management easier, helping you achieve better results.

Other Guides