Study guide: Estimation and Confidence Intervals

Course Outline

  1. Parameters and Sampling Distributions
  2. Sampling Distribution of the Mean
  3. Sampling Distribution of a Proportion
  4. Sampling Distributions for Other Statistics
  5. Point and Interval Estimation
  6. Confidence Level and Interval Structure
  7. Intervals for Population Means
  8. Intervals for Population Proportions
  9. Margin of Error and Sample Size

1. Parameters and Sampling Distributions

Key Concepts & Definitions

  • Parameter: A numerical descriptive measure of a population, based on all population observations and therefore almost always unknown.
  • Sample statistic: A numerical descriptive measure of a sample, calculated from the observations in that sample.
  • Sampling distribution: The probability distribution of a statistic calculated from samples of size n, across those samples.

Essential Points

  • Sampling distributions are used to evaluate sample statistics’ properties before using them to make inferences about population parameters.

Memory Hook

A parameter describes the population; a statistic describes the sample.

2. Sampling Distribution of the Mean

Key Concepts & Definitions

  • Sampling distribution for the mean: The distribution of sample means from all samples of size n.
  • Central Limit Theorem: The Central Limit Theorem states that the mean of a random sample of size n has a sampling distribution that can be approximated by a normal model, with a better approximation for larger n.

★ Must-know

📐 Formula - The mean of the sampling distribution for the sample mean is E(xˉ)=μE(\bar{x})=\mu.

📐 Formula - The standard deviation of the sampling distribution for the sample mean, also called the standard error, is SD(xˉ)=σnSD(\bar{x})=\frac{\sigma}{\sqrt{n}}.

📌 If the population variable is normally distributed, the sampling distribution for the mean is normal; otherwise, it tends toward normality as n increases.

Further detail

  • A sample size of n≥30 is usually enough for the Central Limit Theorem approximation, although the required size depends on the shape of the sampled distribution.

  • For 18-year-old women with population mean weight 143.74 pounds and standard deviation 51.54 pounds, samples of 200 have a mean’s standard error of approximately 3.64 pounds.

  • For those samples of 200 women, a sample mean of 140 pounds has z≈−1.03 and a lower-tail probability of 15.15%, so it is low but not especially unusual.

📐 Formula - For normally distributed heights with population mean 172 cm and standard deviation 10 cm, the sampling distribution for samples of size 50 has standard error 1050≈1.41 cm\frac{10}{\sqrt{50}}\approx1.41\text{ cm}.

  • For the height distribution with mean 172 cm and standard deviation 10 cm, the middle 95% of sample means for samples of size 50 is approximately 169.2 cm to 174.8 cm using the empirical rule.

3. Sampling Distribution of a Proportion

Key Concepts & Definitions

  • Sampling distribution for the proportion: The distribution of sample proportions from all samples of size n.

★ Must-know

📐 Formula - The mean of the sampling distribution for the sample proportion is E(p^)=pE(\hat{p})=p.

📌 The sampling distribution for a proportion is approximately normal for large samples when np≥15np\ge15 and n(1−p)≥15n(1-p)\ge15.

Further detail

  • For a population proportion p, the complementary proportion is denoted q and equals 1−p.

  • For a budget supported by 52% of voters and a poll of 400 voters, np=208 and n(1−p)=192, so the normal model is appropriate.

  • For that poll, the sample proportion has standard error 0.025, and the probability of predicting defeat by observing less than 50% support is approximately 21%.

  • When 13% of people are left-handed, assessing whether 15 left-handed seats suffice for 90 students requires evaluating the probability that the sample proportion exceeds 15/90≈0.167.

4. Sampling Distributions for Other Statistics

Essential Points

  • A sampling distribution can be established for any statistic, but a normal model does not necessarily fit every statistic’s sampling distribution.

5. Point and Interval Estimation

Key Concepts & Definitions

  • Target parameter: The unknown population parameter, such as a mean or proportion, that we want to estimate.
  • Point estimator: A rule or formula that uses sample data to calculate a single number as an estimate of a population parameter.
  • Interval estimator: A formula that uses sample data to calculate a range of values estimating the target parameter.

Essential Points

📌 A mean or average is quantitative and is denoted by μ, whereas a proportion, percentage, fraction, or rate is qualitative and is denoted by p.

  • A point estimate is based on observations from one random sample and does not indicate how close it is to the unknown population parameter.

  • A sample of 200 ESCP students with a mean monthly food expenditure of 364€ gives a point estimate of the unknown population mean.

  • An interval estimator accounts for variation in sample statistics across samples and is stated with a confidence level that is less than 100%.

  • For a sample of 200 ESCP students, a 95% confidence interval for mean monthly food expenditure is 359€ to 369€.

Memory Hook

A point is one mark; an interval is a range that reflects sampling variation.

6. Confidence Level and Interval Structure

Key Concepts & Definitions

  • Confidence level: The probability that a randomly selected confidence interval encloses the population parameter, expressed as a percentage below 100%.
  • Critical value: A table value determined by the sampling distribution of the point estimate and the desired confidence level.
  • Standard error: The standard deviation of the point estimate and measures its variation from sample to sample.

★ Must-know

  • The general confidence-interval formula is Point Estimate±(Critical Value)(Standard Error)\text{Point Estimate} \pm (\text{Critical Value})(\text{Standard Error}).

Further detail

  • The width of the interval from 359€ to 369€ is 10€, equal to twice the distance of 5€ from the point estimate to either limit.

  • A 95% confidence level is the confidence coefficient 1−α=0.951-\alpha=0.95, with α=0.05\alpha=0.05.

  • In repeated sampling, 95% of intervals constructed by this method contain the true parameter and 5% do not.

  • After selecting one sample, the population parameter is unknown, so it cannot be checked whether that sample's interval contains the parameter.

  • Common confidence levels are 80%, 90%, 95%, 98%, 99%, 99.8%, and 99.9%, with corresponding Normal critical values 1.28, 1.645, 1.96, 2.33, 2.576, 3.08, and 3.27.

7. Intervals for Population Means

★ Must-know

📌 For a confidence interval for a population mean with known population standard deviation σ, the sample must be random and have size n≥30n\ge30 to ensure an approximately Normal sampling distribution.

📐 Formula - With known σ, the confidence interval for μ is xˉ±zα/2σn\bar{x} \pm z_{\alpha/2}\frac{\sigma}{\sqrt{n}}.

📌 When σ is unknown, a t-interval requires a random sample and an approximately Normal distribution for the variable.

📐 Formula - With unknown σ, the confidence interval for μ is xˉ±tα/2sn\bar{x} \pm t_{\alpha/2}\frac{s}{\sqrt{n}}.

Further detail

  • For 41 circuits with mean resistance 2.20 ohms and known σ = 0.35 ohms, the 95% confidence interval for mean resistance is [2.09; 2.31] ohms.

  • When σ is unknown, the sample standard deviation s replaces σ, adding uncertainty because s varies from sample to sample, so the Student t-distribution is used.

  • For the t-interval for a mean, the degrees of freedom are n−1n-1, and the critical value depends on the degrees of freedom.

  • The t-distributions are symmetric and bell-shaped with fatter tails than the Normal distribution, and approach the Normal distribution as n increases.

  • For 52 adults, a sample mean pulse rate of 72.7 bpm and standard error of about 0.89 bpm give a 95% t-interval of 70.91 to 74.48 bpm using 51 degrees of freedom and a critical value of about 2.009.

📌 When σ is unknown but the sample is large enough, specified as n≥30n\ge30, t may be replaced by z and the interval is approximately xˉ±zα/2sn\bar{x} \pm z_{\alpha/2}\frac{s}{\sqrt{n}}.

Memory Hook

Known σ uses z; unknown σ uses the t-distribution.

8. Intervals for Population Proportions

★ Must-know

📌 A confidence interval for a population proportion requires a random sample and a sufficiently large sample, with np^≥15n\hat{p}\ge15 and n(1−p^)≥15n(1-\hat{p})\ge15 for an approximately Normal sampling distribution.

📐 Formula - The confidence interval for a population proportion is p^±zα/2p^(1−p^)n\hat{p} \pm z_{\alpha/2}\sqrt{\frac{\hat{p}(1-\hat{p})}{n}}.

Further detail

  • In April 2010, an IPSOS survey of 1011 voters found that 53% were satisfied with Bertrand Delanoë's performance; the 95% confidence interval for the true satisfaction proportion was [49.9%; 56.1%].

  • For the Delanoë survey, the large-sample checks gave 1011(0.53)=536≥151011(0.53)=536\ge15 and 1011(0.47)=475≥151011(0.47)=475\ge15.

9. Margin of Error and Sample Size

Key Concepts & Definitions

  • Margin of error: The amount added to and subtracted from the point estimate to form a confidence interval, and it represents the estimate's imprecision.

★ Must-know

📌 Increasing confidence requires a wider interval, which increases certainty but reduces precision.

Further detail

📐 Formula - A confidence interval can be written as Point Estimate±ME\text{Point Estimate} \pm \text{ME}, where ME=(Critical Value)(Standard Error)\text{ME}=(\text{Critical Value})(\text{Standard Error}).

  • A required sample size can be determined to achieve a specified margin of error at a specified confidence level.

Memory Hook

More confidence requires a wider interval, reducing precision.

Test yourself

Take a 24-question practice quiz on Estimation and Confidence Intervals, with every answer explained.

1. A researcher wants to describe the average income of every resident in a city, but has data from the entire population. What kind of quantity is that average?

2. An analyst calculates the average satisfaction rating from the 80 customers who responded to a survey. What is this average?

Take the quiz →

Study with flashcards

Memorize the key concepts of Estimation and Confidence Intervals with 59 flashcards.

What is a parameter in population statistics?

A numerical descriptive measure of a population based on all its observations.

Why is a population parameter almost always unknown?

It is based on all population observations.

What is a sample statistic?

A numerical descriptive measure calculated from a sample’s observations.

See the flashcards →

Related classes

Make your own study guides

Upload your class notes and Revizly writes a study guide, a quiz and flashcards in about a minute.

Study guide maker