If you're seeing this message, it means we're having trouble loading external resources on our website.

If you're behind a web filter, please make sure that the domains *.kastatic.org and *.kasandbox.org are unblocked.

# Interpreting a confidence interval for a mean

After we build a confidence interval for a mean, it's important to be able to interpret what the interval tells us about the population and what it doesn't tell us.
A confidence interval for a mean gives us a range of plausible values for the population mean. If a confidence interval does not include a particular value, we can say that it is not likely that the particular value is the true population mean. However, even if a particular value is within the interval, we shouldn't conclude that the population mean equals that specific value.
Let's look at few examples that demonstrate how to interpret a confidence interval for a mean.

## Example 1

Felix is a quality control expert at a factory that paints car parts. Their painting process consists of a primer coat, color coat, and clear coat. For a certain part, these layers have a combined target thickness of $150$ microns. Felix measured the thickness of $50$ randomly selected points on one of these parts to see if it was painted properly. His sample had a mean thickness of $\overline{x}=148$ microns and a standard deviation of ${s}_{x}=3.3$ microns.
A $95\mathrm{%}$ confidence interval for the mean thickness based on his data is $\left(147.1,148.9\right)$.
Based on his interval, is it plausible that this part's average thickness agrees with the target value?
No, it isn't. The interval says that the plausible values for the true mean thickness on this part are between $147.1$ and $148.9$ microns. Since this interval doesn't contain $150$ microns, it doesn't seem plausible that this part's average thickness agrees with the target value. In other words, the entire interval is below the target value of $150$ microns, so this part's mean thickness is likely below the target.

## Example 2

Martina read that the average graduate student is $33$ years old. She wanted to estimate the mean age of graduate students at her large university, so she took a random sample of $30$ graduate students. She found that their mean age was $\overline{x}=31.8$ and the standard deviation was ${s}_{x}=4.3$ years. A $95\mathrm{%}$ confidence interval for the mean based on her data was $\left(30.2,33.4\right)$.
Based on this interval, is it plausible that the mean age of all graduate students at her university is also $33$ years?
Yes. Since $33$ is within the interval, it is a plausible value for the mean age of the entire population of graduate students at her university.

## Example 3: Try it out!

The Environmental Protection Agency (EPA) has standards and regulations that say that the lead level in soil cannot exceed the limit of $400$ parts per million (ppm) in public play areas designed for children. Luke is an inspector, and he takes $30$ randomly selected soil samples from a site where they are considering building a playground.
These data show a sample mean of $\overline{x}=394\phantom{\rule{0.167em}{0ex}}\text{ppm}$ and a standard deviation of ${s}_{x}=26.3\phantom{\rule{0.167em}{0ex}}\text{ppm}$. The resulting $95\mathrm{%}$ confidence interval for the mean lead level is $394±9.8.$
What does this interval suggest?

## Example 4: Try it out!

Sandra is an engineer working on wireless charging for a mobile phone manufacturer. Their design specifications say that it should take no more than $2$ hours to completely charge a fully depleted battery.
Sandra took a random sample of $40$ of these phones and chargers. She fully depleted their batteries and timed how long it took each of them to completely charge. Those measurements were used to construct a $95\mathrm{%}$ confidence interval for the mean charging time. The resulting interval was $124±2.24$ minutes.
What does this interval suggest about the charging times?
Hint: The specification was "no more than $2$ hours."

## Want to join the conversation?

• The wording on Example 1 is a bit confusing: "For a certain part, these layers have a combined target thickness of 150 microns. Felix measured the thickness of 50 randomly selected points on one of these parts to see if it was painted properly."

In the first sentence I quoted, does "a certain part" describe the population mean across all instances of that "certain part"? Or, is it talking about the thickness of a specific instance of that part? In other words, the specificity of "certain part" is ambiguous.

Then in the second sentence it says "one of these parts". But, it is ambiguous what "these parts" is referring to. Is it referring to an instance of the type of part from the previous sentence, or is it generally talking about some part of the many parts the factory manufactures.

I would suggest using more specific words to make this more clear. Maybe: "The mean thickness on any point of any Part #1234 is 150 microns. Felix takes a single instance of Part #1234 to inspect and measures it in 50 randomly selected spots."

I'm not sure how to fix it, but it is very hard to read as it can be interpreted in several ways.
• I think the problem is considering the population to be all the points in each one of the pieces manufactured. That is the only way I can make sense of it.
(1 vote)
• a random sample of 36 drivers used on average 749 gallons of gasoline per year. if the standard deviation of the population is 32 gallons, find the 95% confidence interval of the mean for all drivers. if a driver said that he used 803 gallons per year, would you believe that?
(1 vote)
• The conditions for inference are met and so the confidence interval is
𝑥̅ ± 𝑧* ∙ 𝜎∕√𝑛 =
= 749 ± 1.96 ∙ 32∕√36 ≈
≈ (738, 760)

This means that we are 95% confident that the population mean is within this interval.

It doesn't tell us anything about the shape of the population distribution though.
Thereby we can't really tell how likely it is for someone to use 803 gallons of fuel per year and without other evidence there is no reason for us to doubt a driver who claims they use that much fuel.
• So, like, I still have trouble with this... Where do we even get the 95, 90, 99% confidence? Do we just make it up? Just choose randomly, "I feel like a 97% today..." or what?
• It's a convention. You would be free to use what figure you like but it is commonplace/traditional to use a stat of 95% or 99%.
• In the examples where the population mean is in the confidence interval, do we not need to take multiple sets of samples to conclude that it is in the interval 95% of the time?
(1 vote)
• you could do that to prove that the formulas work.. But it has been proven and you can safely assume the formulas work. Keep in mind you need to meet the requirements for your sample: randomness, normality and independence.
• Are all plausible values in the confidence interval equally likely. I am inclined to believe that the values in the middle of the confidence interval are more likely to be the true mean than those at the extremes. Since, those values outside of the confidence interval are considered not plausible, for instance if we question whether 44.9 or 55.1 are plausible true means in a sample giving confidence interval for mean as 45-55, the answer will be no. So, if 44.9 is not plausible and 55.1 is also not plausible, then logically 44 or 55 may not suddenly become plausible with the same likelihood for the entire interval. I am inclined to believe that they will be slightly more plausible than those values just outside the confidence interval, and the plausibility of values will increase as we approach the middle of the confidence interval. Can someone clarify this please? Thanks.
(1 vote)
• Plausible values within the confidence interval are not equally likely. The values closer to the center of the interval are more plausible than those at the extremes because the interval reflects a range of values within which the true population parameter is likely to lie. However, the likelihood of each specific value within the interval is not uniform.
(1 vote)
• I'm a bit confused. In previous lessons, it was specified that the confidence interval gave the likelihood (in%) that the real mean is contained in (x%) of the trials over numerous trials. Now it's saying the percentage gives you the likelihood that a specific interval contains the true percentage?
(1 vote)
• The confidence interval does not give the likelihood that a specific interval contains the true parameter value. Instead, it provides a range of values within which the true parameter is likely to lie with a certain level of confidence. The confidence level represents the proportion of intervals, constructed in the same way from the same population, that would contain the true parameter value if the process were repeated indefinitely.
(1 vote)
• Is it still plausible if the claim/desired value falls at the endpoint of a given confidence interval?
(1 vote)
• If the claim or desired value falls exactly at the endpoint of a given confidence interval, it means that there is uncertainty about whether the true parameter value lies exactly at that point. The interval provides a range of plausible values, and the claim falls within that range. However, it does not necessarily imply that the claim is proven or guaranteed to be true.
(1 vote)
• Would we still be able to make the same conclusions if the confidence level were lower, for example 90%? I thought it would technically no longer meet the 0.05 default significance level, but I was told that the confidence level isn't too important and that we would still make the same conclusions?

Any last-minute insight before the ap exam would be greatly appreciated 😅
(1 vote)
• It's too bad that I didn't get to this before the exam came around! I hope that it went well for you.

The confidence level is important because it could work in conjunction with the idea of power and Type I/Type II errors.
(1 vote)
• a study was done to see if there is a difference between the number of sick days men take and the number of sick days women take. a random sample of 9 men found that the mean of the number of sick days taken was 5.5. The standard deviation of the sample was 1.23. a random sample of 7 women found that the mean was 4.3 days and a standard deviation of 1.19 days. at alpha = 0.05, can it be concluded that there is a difference in the means?
• To test for a difference in means, calculate the pooled standard deviation:
pooled standard deviation = sqrt[((n1 - 1)*s1^2 + (n2 - 1)*s2^2) / (n1 + n2 - 2)]
pooled standard deviation ≈ sqrt[((9 - 1)*1.23^2 + (7 - 1)*1.19^2) / (9 + 7 - 2)] ≈ 1.213

Then, calculate the t-statistic:
t = (mean1 - mean2) / (pooled standard deviation * sqrt(1/n1 + 1/n2))
t ≈ (5.5 - 4.3) / (1.213 * sqrt(1/9 + 1/7)) ≈ 2.08

Finally, compare the t-statistic to the critical value from the t-distribution with (n1 + n2 - 2) degrees of freedom at α = 0.05.
Since the calculated t-value (2.08) > critical t-value, reject the null hypothesis. There is evidence to suggest a difference in means.
(1 vote)
• in order to monitor the weight of cows 6 randomly chosen the mean is -8.00 and standard deviation 2.828 construct 95% confidence interval