Skip to main content
Statistics LibreTexts

2.2: Confidence Intervals

  • Page ID
    2877
  • \( \newcommand{\vecs}[1]{\overset { \scriptstyle \rightharpoonup} {\mathbf{#1}} } \)

    \( \newcommand{\vecd}[1]{\overset{-\!-\!\rightharpoonup}{\vphantom{a}\smash {#1}}} \)

    \( \newcommand{\id}{\mathrm{id}}\) \( \newcommand{\Span}{\mathrm{span}}\)

    ( \newcommand{\kernel}{\mathrm{null}\,}\) \( \newcommand{\range}{\mathrm{range}\,}\)

    \( \newcommand{\RealPart}{\mathrm{Re}}\) \( \newcommand{\ImaginaryPart}{\mathrm{Im}}\)

    \( \newcommand{\Argument}{\mathrm{Arg}}\) \( \newcommand{\norm}[1]{\| #1 \|}\)

    \( \newcommand{\inner}[2]{\langle #1, #2 \rangle}\)

    \( \newcommand{\Span}{\mathrm{span}}\)

    \( \newcommand{\id}{\mathrm{id}}\)

    \( \newcommand{\Span}{\mathrm{span}}\)

    \( \newcommand{\kernel}{\mathrm{null}\,}\)

    \( \newcommand{\range}{\mathrm{range}\,}\)

    \( \newcommand{\RealPart}{\mathrm{Re}}\)

    \( \newcommand{\ImaginaryPart}{\mathrm{Im}}\)

    \( \newcommand{\Argument}{\mathrm{Arg}}\)

    \( \newcommand{\norm}[1]{\| #1 \|}\)

    \( \newcommand{\inner}[2]{\langle #1, #2 \rangle}\)

    \( \newcommand{\Span}{\mathrm{span}}\) \( \newcommand{\AA}{\unicode[.8,0]{x212B}}\)

    \( \newcommand{\vectorA}[1]{\vec{#1}}      % arrow\)

    \( \newcommand{\vectorAt}[1]{\vec{\text{#1}}}      % arrow\)

    \( \newcommand{\vectorB}[1]{\overset { \scriptstyle \rightharpoonup} {\mathbf{#1}} } \)

    \( \newcommand{\vectorC}[1]{\textbf{#1}} \)

    \( \newcommand{\vectorD}[1]{\overrightarrow{#1}} \)

    \( \newcommand{\vectorDt}[1]{\overrightarrow{\text{#1}}} \)

    \( \newcommand{\vectE}[1]{\overset{-\!-\!\rightharpoonup}{\vphantom{a}\smash{\mathbf {#1}}}} \)

    \( \newcommand{\vecs}[1]{\overset { \scriptstyle \rightharpoonup} {\mathbf{#1}} } \)

    \( \newcommand{\vecd}[1]{\overset{-\!-\!\rightharpoonup}{\vphantom{a}\smash {#1}}} \)

    \(\newcommand{\avec}{\mathbf a}\) \(\newcommand{\bvec}{\mathbf b}\) \(\newcommand{\cvec}{\mathbf c}\) \(\newcommand{\dvec}{\mathbf d}\) \(\newcommand{\dtil}{\widetilde{\mathbf d}}\) \(\newcommand{\evec}{\mathbf e}\) \(\newcommand{\fvec}{\mathbf f}\) \(\newcommand{\nvec}{\mathbf n}\) \(\newcommand{\pvec}{\mathbf p}\) \(\newcommand{\qvec}{\mathbf q}\) \(\newcommand{\svec}{\mathbf s}\) \(\newcommand{\tvec}{\mathbf t}\) \(\newcommand{\uvec}{\mathbf u}\) \(\newcommand{\vvec}{\mathbf v}\) \(\newcommand{\wvec}{\mathbf w}\) \(\newcommand{\xvec}{\mathbf x}\) \(\newcommand{\yvec}{\mathbf y}\) \(\newcommand{\zvec}{\mathbf z}\) \(\newcommand{\rvec}{\mathbf r}\) \(\newcommand{\mvec}{\mathbf m}\) \(\newcommand{\zerovec}{\mathbf 0}\) \(\newcommand{\onevec}{\mathbf 1}\) \(\newcommand{\real}{\mathbb R}\) \(\newcommand{\twovec}[2]{\left[\begin{array}{r}#1 \\ #2 \end{array}\right]}\) \(\newcommand{\ctwovec}[2]{\left[\begin{array}{c}#1 \\ #2 \end{array}\right]}\) \(\newcommand{\threevec}[3]{\left[\begin{array}{r}#1 \\ #2 \\ #3 \end{array}\right]}\) \(\newcommand{\cthreevec}[3]{\left[\begin{array}{c}#1 \\ #2 \\ #3 \end{array}\right]}\) \(\newcommand{\fourvec}[4]{\left[\begin{array}{r}#1 \\ #2 \\ #3 \\ #4 \end{array}\right]}\) \(\newcommand{\cfourvec}[4]{\left[\begin{array}{c}#1 \\ #2 \\ #3 \\ #4 \end{array}\right]}\) \(\newcommand{\fivevec}[5]{\left[\begin{array}{r}#1 \\ #2 \\ #3 \\ #4 \\ #5 \\ \end{array}\right]}\) \(\newcommand{\cfivevec}[5]{\left[\begin{array}{c}#1 \\ #2 \\ #3 \\ #4 \\ #5 \\ \end{array}\right]}\) \(\newcommand{\mattwo}[4]{\left[\begin{array}{rr}#1 \amp #2 \\ #3 \amp #4 \\ \end{array}\right]}\) \(\newcommand{\laspan}[1]{\text{Span}\{#1\}}\) \(\newcommand{\bcal}{\cal B}\) \(\newcommand{\ccal}{\cal C}\) \(\newcommand{\scal}{\cal S}\) \(\newcommand{\wcal}{\cal W}\) \(\newcommand{\ecal}{\cal E}\) \(\newcommand{\coords}[2]{\left\{#1\right\}_{#2}}\) \(\newcommand{\gray}[1]{\color{gray}{#1}}\) \(\newcommand{\lgray}[1]{\color{lightgray}{#1}}\) \(\newcommand{\rank}{\operatorname{rank}}\) \(\newcommand{\row}{\text{Row}}\) \(\newcommand{\col}{\text{Col}}\) \(\renewcommand{\row}{\text{Row}}\) \(\newcommand{\nul}{\text{Nul}}\) \(\newcommand{\var}{\text{Var}}\) \(\newcommand{\corr}{\text{corr}}\) \(\newcommand{\len}[1]{\left|#1\right|}\) \(\newcommand{\bbar}{\overline{\bvec}}\) \(\newcommand{\bhat}{\widehat{\bvec}}\) \(\newcommand{\bperp}{\bvec^\perp}\) \(\newcommand{\xhat}{\widehat{\xvec}}\) \(\newcommand{\vhat}{\widehat{\vvec}}\) \(\newcommand{\uhat}{\widehat{\uvec}}\) \(\newcommand{\what}{\widehat{\wvec}}\) \(\newcommand{\Sighat}{\widehat{\Sigma}}\) \(\newcommand{\lt}{<}\) \(\newcommand{\gt}{>}\) \(\newcommand{\amp}{&}\) \(\definecolor{fillinmathshade}{gray}{0.9}\)

    In the preceding chapter we learned that populations are characterized by descriptive measures called parameters. Inferences about parameters are based on sample statistics. We now want to estimate population parameters and assess the reliability of our estimates based on our knowledge of the sampling distributions of these statistics.

    Point Estimates

    We start with a point estimate. This is a single value computed from the sample data that is used to estimate the population parameter of interest.

    • The sample mean (\(\bar {x}\)) is a point estimate of the population mean (\(\mu\)).
    • The sample proportion (\(\hat {p}\)) is the point estimate of the population proportion (p).

    We use point estimates to construct confidence intervals for unknown parameters.

    • A confidence interval is an interval of values instead of a single point estimate.
    • The level of confidence corresponds to the expected proportion of intervals that will contain the parameter if many confidence intervals are constructed of the same sample size from the same population.
    • Our uncertainty is about whether our particular confidence interval is one of those that truly contains the true value of the parameter.
    Example \(\PageIndex{1}\): bear weight

    We are 95% confident that our interval contains the population mean bear weight.

    If we created 100 confidence intervals of the same size from the same population, we would expect 95 of them to contain the true parameter (the population mean weight). We also expect five of the intervals would not contain the parameter.

    imageedit_4_4261276457.png

    Figure \(\PageIndex{1}\): Confidence intervals from twenty-five different samples.

    In this example, twenty-five samples from the same population gave these 95% confidence intervals. In the long term, 95% of all samples give an interval that contains µ, the true (but unknown) population mean.

    Level of confidence is expressed as a percent.

    • The compliment to the level of confidence is α (alpha), the level of significance.
    • The level of confidence is described as \((1- \alpha) \times 100%\).

    What does this really mean?

    • We use a point estimate (e.g., sample mean) to estimate the population mean.
    • We attach a level of confidence to this interval to describe how certain we are that this interval actually contains the unknown population parameter.
    • We want to estimate the population parameter, such as the mean (μ) or proportion (p).

    \[\bar {x}-E < \mu < \bar {x}+E\]

    or

    \[\hat {p}-E < p <\hat {p}+E\]

    where \(E\) is the margin of error.

    The confidence is based on area under a normal curve. So the assumption of normality must be met (Chapter 1).

    Confidence Intervals about the Mean (μ) when the Population Standard Deviation (σ) is Known

    A confidence interval takes the form of: point estimate \(\pm\) margin of error.

    The point estimate

    • The point estimate comes from the sample data.
    • To estimate the population mean (\(μ\)), use the sample mean (\(\bar{x}\)) as the point estimate.

    The margin of error

    • Depends on the level of confidence, the sample size and the population standard deviation.
    • It is computed as \(E=Z_{\frac {\alpha}{2}}\times \frac {\sigma}{\sqrt {n}}\)where \(Z_{\frac {\alpha}{2}}\) is the critical value from the standard normal table associated with α (the level of significance).

    The critical value \(Z_{\frac {\alpha}{2}}\)

    • This is a Z-score that bounds the level of confidence.
    • Confidence intervals are ALWAYS two-sided and the Z-scores are the limits of the area associated with the level of confidence.
    clipboard_e3b241bc8de6aeefb18ca6be09857390c.png
    Figure \(\PageIndex{1}\): The middle 95% area under a standard normal curve.
    • The level of significance (α) is divided into halves because we are looking at the middle 95% of the area under the curve.
    • Go to your standard normal table and find the area of 0.025 in the body of values.
    • What is the Z-score for that area?

    Table \(\PageIndex{1}\): Common critical values (Z-scores).

    Confidence Level

    (level of significance)

    \(Z_{\alpha / 2}\)

    99%

    1%

    2.575

    95%

    5%

    1.96

    90%

    10%

    1.645

    Steps

    Construction of a confidence interval about \(μ\) when \(σ\) is known:

    1. \(Z_{\frac {\alpha}{2}}\) (critical value)
    2. \(E=Z_{\frac {\alpha}{2}}\times \frac {\sigma}{\sqrt {n}}\) (margin of error)
    3. \(\bar {x} \pm E\) (point estimate ± margin of error)
    Example \(\PageIndex{3}\): Construct a confidence interval about the population mean

    Researchers have been studying p-loading in Jones Lake for many years. It is known that mean water clarity (using a Secchi disk) is normally distributed with a population standard deviation of σ = 15.4 in. A random sample of 22 measurements was taken at various points on the lake with a sample mean of = 57.8 in. The researchers want you to construct a 95% confidence interval for μ, the mean water clarity.

    imageedit_7_3807469364.png

    A secchi disk to measure turbidly of water. (CC SA; publiclab.org)

    Solution

    1) \(Z_{\frac {\alpha}{2}}\) = 1.96

    2) \(E=Z_{\frac {\alpha}{2}}\times \frac {\sigma}{\sqrt {n}}\) = \(1.96 \times \frac {15.4}{\sqrt {22}}\) = 6.435

    3) \(\bar {x} \pm E\) = 57.8 ± 6.435

    95% confidence interval for the mean water clarity is (51.36, 64.24).

    We can be 95% confident that this interval contains the population mean water clarity for Jones Lake.

    Now construct a 99% confidence interval for μ, the mean water clarity, and interpret.

    1) \(Z_{\frac {\alpha}{2}}\)= 2.575

    2) \(E=Z_{\frac {\alpha}{2}}\times \frac {\sigma}{\sqrt {n}}\) = \(2.575 \times \frac {15.4}{\sqrt {22}}\) = 8.454

    3) \(\bar {x} \pm E\)= 57.8± 8.454

    99% confidence interval for the mean water clarity is (49.35, 66.25).

    We can be 99% confident that this interval contains the population mean water clarity for Jones Lake.

    As the level of confidence increased from 95% to 99%, the width of the interval increased. As the probability (area under the normal curve) increased, the critical value increased resulting in a wider interval.

    Software Solutions

    Minitab

    You can use Minitab to construct this 95% confidence interval (Excel does not construct confidence intervals about the mean when the population standard deviation is known). Select Basic Statistics>1-sample Z. Enter the known population standard deviation and select the required level of confidence.

    030_2.tif

    030_1.tif

    Minitab screen shots for constructing a confidence interval.

    One-Sample Z: depth

    The assumed standard deviation = 15.4

    Variable

    N

    Mean

    StDev

    SE Mean

    95% CI

    depth

    22

    57.80

    11.60

    3.28

    (51.36, 64.24)

    Confidence Intervals about the Mean (μ) when the Population Standard Deviation (σ) is Unknown

    Typically, in real life we often don’t know the population standard deviation (σ). We can use the sample standard deviation (s) in place of σ. However, because of this change, we can’t use the standard normal distribution to find the critical values necessary for constructing a confidence interval.

    The Student’s t-distribution was created for situations when σ was unknown. Gosset worked as a quality control engineer for Guinness Brewery in Dublin. He found errors in his testing and he knew it was due to the use of s instead of σ. He created this distribution to deal with the problem of an unknown population standard deviation and small sample sizes. A portion of the t-table is shown below.

    Table \(\PageIndex{2}\): Portion of the student’s t-table.

    df

    0.10

    0.05

    0.025

    0.02

    0.01

    0.005

    1

    3.078

    6.314

    12.706

    15.894

    31.821

    63.657

    2

    1.886

    2.920

    4.303

    4.849

    6.965

    9.925

    3

    1.638

    2.353

    3.182

    3.482

    4.541

    5.841

    4

    1.533

    2.132

    2.776

    2.999

    3.747

    4.604

    5

    1.476

    2.015

    2.571

    2.757

    3.365

    4.

    Example \(\PageIndex{4}\)

    Find the critical value \(t_{\frac {\alpha}{2}}\) for a 95% confidence interval with a sample size of n=13.

    Solution

    • Degrees of freedom (down the left-hand column) is equal to n-1 = 12
    • α = 0.05 and α/2 = 0.025
    • Go down the 0.025 column to 12 df
    • \(t_{\frac {\alpha}{2}}\)= 2.179

    The critical values from the students’ t-distribution approach the critical values from the standard normal distribution as the sample size (n) increases.

    Table \(\PageIndex{3}\): Critical values from the student’s t-table.

    n

    Degrees of Freedom

    \(t_{0.25}\)

    11

    10

    2.228

    51

    50

    2.009

    101

    100

    1.984

    1001

    1000

    1.962

    Using the standard normal curve, the critical value for a 95% confidence interval is 1.96. You can see how different samples sizes will change the critical value and thus the confidence interval, especially when the sample size is small.

    Construction of a Confidence Interval

    When σ is Unknown

    1. \(t_{\frac {\alpha}{2}}\) critical value with n-1 df
    2. \(E = t_{\frac {\alpha}{2}} \times \frac{s}{\sqrt {n}}\)
    3. \(\bar {x} \pm E\)
    Example \(\PageIndex{5}\):

    Researchers studying the effects of acid rain in the Adirondack Mountains collected water samples from 22 lakes. They measured the pH (acidity) of the water and want to construct a 99% confidence interval about the mean lake pH for this region. The sample mean is 6.4438 with a sample standard deviation of 0.7120. They do not know anything about the distribution of the pH of this population, and the sample is small (n<30), so they look at a normal probability plot.

    2970.png

    Figure 4. Normal probability plot.

    Solution

    The data is normally distributed. Now construct the 99% confidence interval about the mean pH.

    1) \(t_{\frac {\alpha}{2}}\) = 2.831

    2) \(E = t_{\frac {\alpha}{2}} \times \frac{s}{\sqrt {n}}\) = \(2.831 \times \frac {0.7120}{\sqrt {22}}\)= 0.4297

    3) \(\bar {x} \pm E\) = 6.443 ± 0.4297

    The 99% confidence interval about the mean pH is (6.013, 6.863).

    We are 99% confident that this interval contains the mean lake pH for this lake population.

    Now construct a 90% confidence interval about the mean pH for these lakes.

    1) \(t_{\frac {\alpha}{2}}\) = 1.721

    2) \(E = t_{\frac {\alpha}{2}} \times \frac{s}{\sqrt {n}}\) = \(1.71221 \times \frac {0.7120}{\sqrt {22}}\)0.2612

    3) \(\bar {x} \pm E\) = 6.443 ± 0.2612

    The 90% confidence interval about the mean pH is (6.182, 6.704).

    We are 90% confident that this interval contains the mean lake pH for this lake population.

    Notice how the width of the interval decreased as the level of confidence decreased from 99 to 90%.

    Construct a 90% confidence interval about the mean lake pH using Excel and Minitab.

    Software Solutions

    Minitab

    For Minitab, enter the data in the spreadsheet and select Basic statistics and 1-sample t-test.

    035_2.tif

    035_1.tif

    One-Sample T: pH

    Variable N Mean StDev SE Mean 90% CI
    pH

    22

    6.443

    0.712

    0.152

    (6.182, 6.704)

    Additional example:

    Excel

    For Excel, enter the data in the spreadsheet and select descriptive statistics. Check Summary Statistics and select the level and confidence.

    034_2.tif

    034_1.tif

    Mean

    6.442909

    Standard Error

    0.151801

    Median

    6.4925

    Mode

    #N/A

    Standard Deviation

    0.712008

    Sample Variance

    0.506956

    Kurtosis

    -0.5007

    Skewness

    -0.60591

    Range

    2.338

    Minimum

    5.113

    Maximum

    7.451

    Sum

    141.744

    Count

    22

    Confidence Level(90.0%)

    0.26121

    Excel gives you the sample mean in the first line (6.442909) and the margin of error in the last line (0.26121). You must complete the computation yourself to obtain the interval (6.442909±0.26121).

    Confidence Intervals about the Population Proportion (p)

    Frequently, we are interested in estimating the population proportion (p), instead of the population mean (µ). For example, you may need to estimate the proportion of trees infected with beech bark disease, or the proportion of people who support “green” products. The parameter p can be estimated in the same ways as we estimated µ, the population mean.

    The Sample Proportion

    • The sample proportion is the best point estimate for the true population proportion.
    • Sample proportion \(\hat {p} = \frac {x}{n}\)where x is the number of elements in the sample with the characteristic you are interested in, and n is the sample size.

    The Assumption of Normality when Estimating Proportions

    • The assumption of a normally distributed population is still important, even though the parameter has changed.
    • Normality can be verified if:$$ n \times \hat {p} \times (1- \hat {p}) \ge 10$$

    Constructing a Confidence Interval about the Population Proportion

    Constructing a confidence interval about the proportion follows the same three steps we have used in previous examples.

    1. \(Z_{\frac {\alpha}{2}}\)(critical value from the standard normal table)
    2. \(E = Z_{\frac {\alpha}{2}} \times \sqrt {\frac{\hat {p}(1-\hat {p})}{n}}\) (margin of error)
    3. \(\hat {p} \pm E\)(point estimate ± margin of error)
    Example \(\PageIndex{6}\):

    A botanist has produced a new variety of hybrid soybean that is better able to withstand drought. She wants to construct a 95% confidence interval about the germination rate (percent germination). She randomly selected 500 seeds and found that 421 have germinated.

    Solution

    First, compute the point estimate

    \[\hat {p} = \frac {x}{n} =\frac {421}{500}=0.842\]

    Check normality:

    \[n \times \hat {p} \times (1-\hat {p}) \ge 10 = 500 \times 0.842 \times (1-0.842) =66.5\]

    You can assume a normal distribution.

    Now construct the confidence interval:

    1) \(Z_{\frac {\alpha}{2}}\) = 1.96

    2) \(E = Z_{\frac {\alpha}{2}} \times \sqrt {\frac{\hat {p}(1-\hat {p})}{n}}\) =\(1.96 \times \sqrt {\frac {0.842(1-0.842)}{500}}\) = 0.032

    3) \(\hat {p} \pm E =0.842 \pm 0.0032\)

    The 95% confidence interval for the germination rate is (81.0%, 87.4%).

    We can be 95% confident that this interval contains the true germination rate for this population.

    Software Solutions

    Minitab

    You can use Minitab to compute the confidence interval. Select STAT>Basic stats>1-proportion. Select summarized data and enter the number of events (421) and the number of trials (500). Click Options and select the correct confidence level. Check “test and interval based on normal distribution” if the assumption of normality has been verified.

    036_1.tif

    036_2.tif

    Test and CI for One Proportion

    Sample X N Sample p 95% CI
    1 421 500 0.842000 (0.810030, 0.873970)

    Using the normal approximation.

    Excel

    Excel does not compute confidence intervals for estimating the population proportion.

    Confidence Interval Summary

    Which method do I use?

    The first question to ask yourself is: Which parameter are you trying to estimate? If it is the mean (µ), then ask yourself: Is the population standard deviation (σ) known? If yes, then follow the next 3 steps:

    Confidence Interval about the Population Mean (µ) when σ is Known

    1. \(Z_{\frac {\alpha}{2}}\) critical value (from the standard normal table)
    2. \(E=Z_{\frac {\alpha}{2}} \times \frac {\sigma}{\sqrt {n}}\)
    3. \(\bar {x} \pm E\)

    If no, follow these 3 steps:

    Confidence Interval about the Population Mean (µ) when σ is Unknown

    1. \(t_{\frac {\alpha}{2}}\) critical value with n-1 df from the student t-distribution
    2. \(E=t_{\frac {\alpha}{2}} \times \frac {s}{\sqrt {n}}\)
    3. \(\bar {x} \pm E\)

    If you want to construct a confidence interval about the population proportion, follow these 3 steps:

    Confidence Interval about the Proportion

    1. \(Z_{\frac {\alpha}{2}}\) critical value from the standard normal table
    2. \(E = Z_{\frac {\alpha}{2}} \times \sqrt {\frac{\hat {p}(1-\hat {p})}{n}}\)
    3. \(\hat {p} \pm E\)

    Remember that the assumption of normality must be verified.


    This page titled 2.2: Confidence Intervals is shared under a CC BY-NC-SA 3.0 license and was authored, remixed, and/or curated by Diane Kiernan (OpenSUNY) via source content that was edited to the style and standards of the LibreTexts platform.