← Course materials

EN.553.211 · Fall 2026 · Zan Ahmad

Chapter 7: Confidence Intervals

The current lecture deck has 83 PDF pages. The existing 18-page notes cover the chapter with explanations and additional practice. Sections below follow the slide order; finite-population sampling appears later in the notes. Questions and previously released solutions remain in their existing locations.

Practice and worked solutions

Page numbers refer to the PDF viewer, including repeated slide reveals. Use the links below or the navigation strip in either PDF to switch between notes and slides. Some mobile or embedded viewers open at page 1; if so, enter the page number shown here. Navigation sits above the original page; content, page order and existing page references are unchanged.

November 5 revision: notation follows the course slides; figures connect the equations to their meaning. Notes retain their existing page numbers. On a small screen, swipe a wide figure or open its full-size view.

Parameters, estimators and chapter overview

Finite-population correction and small-population example

The slide deck begins here; the notes place this topic after sample-size planning. PDF page numbers are preserved.

The correction sqrt((N-n)/(N-1)) applies when the finite-population variance uses divisor N; it approaches zero as the sample becomes a census.
Sampling without replacement. The correction sqrt((N-n)/(N-1)) applies when the finite-population variance uses divisor N; it approaches zero as the sample becomes a census. Open full-size figure

From point estimation to interval estimation

Central normal probabilities and critical values

Sampling distribution and confidence interval for one mean

Derivation on notes p. 2; repeated-sampling interpretation on p. 3. A fixed observed interval does not assign a probability to the unknown parameter.

Estimated standard error and large-sample mean interval

Replacing sigma by S gives an approximate large-sample normal procedure, not an exactly normal statistic at n = 30. Notes p. 6 makes this distinction explicit.

Dietitian example

The problem and calculation use n = 100. The first line on slide PDF p. 37 says n = 50; this is a source typo. Notes p. 6 consistently uses n = 100 and obtains [0.70, 14.42] grams/day.

Interpreting confidence

Illustrative repeated samples: the parameter stays fixed while intervals move. This small sketch does not assert an exact observed percentage.
Confidence is long-run coverage. Illustrative repeated samples: the parameter stays fixed while intervals move. This small sketch does not assert an exact observed percentage. Open full-size figure

Confidence level, precision and interval width

Hold confidence and the population variability fixed. Halving the margin of error requires four times the sample size.
Precision scales as one over square-root n. Hold confidence and the population variability fixed. Halving the margin of error requires four times the sample size. Open full-size figure

Why sample variance uses n − 1

Notes p. 6 introduces estimated variability; p. 15 contains the full derivation.

Difference between two independent population means

Estimated-SE normal inference is approximate; adequate sample sizes and independent groups are separate requirements.

Subtracting independent estimates does not cancel their uncertainty. Square the SEs, add, then take the square root.
Independent uncertainty adds in variance. Subtracting independent estimates does not cancel their uncertainty. Square the SEs, add, then take the square root. Open full-size figure

Sampling distribution of a proportion

For fixed n, Bernoulli variability is largest at p = 0.5. This motivates conservative sample-size planning.
Proportion variability. For fixed n, Bernoulli variability is largest at p = 0.5. This motivates conservative sample-size planning. Open full-size figure

Blue-eyes probability example

For p = 0.09 and n = 613, the standard error is 0.011559 and the direct normal tail is about 0.0418. The values 0.0121 and 0.0564 on slide PDF pp. 61–62 are arithmetic errors. Notes p. 8 gives the corrected calculation and distinguishes it from a continuity-corrected count calculation.

Confidence interval for one proportion

Difference between two independent proportions

Bond-proposal example

Choosing a sample size and planning values

The notes explain the limits of the range/4 rule and why p = 1/2 gives conservative proportion planning.

95% confidence with a one-percentage-point margin

Additional Practice 1: one mean

This water-specimen question is additional course practice. Related slide sections: one-mean confidence intervals, interpretation and width.

Formula map, inverse calculations and final checks

These pages synthesize the chapter and add checks; they do not correspond to one slide range. The Chapter 6 sum tools on notes p. 14 connect to Chapter 6 slide PDF pp. 62–81.

Additional Problem 2: two means

This protein-concentration question is additional course practice. Use the two-means section, slide PDF pp. 46–53.

Additional Problem 3: one proportion

This usable-water-sample question is additional course practice. Related slides: sampling proportions pp. 54–62, proportion intervals pp. 63–64 and sample-size planning pp. 72–83.

Office-hours appendix

The existing office-hours appendix is preserved. It has no lecture-slide counterpart.