BUSS1020 Chap.7 Confidence Intervals and Sample Size
Confidence Intervals and Sample Size
Confidence intervals add an explicit uncertainty band to a point estimate. This chapter uses the structure estimate plus or minus critical value times standard error, choosing z or Student's t according to the information available. It covers intervals for means and proportions, correct repeated-sampling interpretation, margin-of-error relationships and sample-size planning.
The practical focus is on stating the parameter and unit, rounding planned sample size upward and keeping survey design limitations visible. Width becomes a decision variable: confidence, variability and sample size each change it in a predictable direction. The worked t interval prints degrees of freedom and checks endpoint symmetry, while the planning drills connect a tolerated margin to fieldwork cost.
A repeated-sampling interpretation avoids assigning probability to a fixed parameter after the interval has been observed. The chapter also separates completed observations from invitations, because non-response changes fieldwork needs without changing the formula's required usable sample. Every endpoint is tied back to a decision tolerance rather than presented as precision for its own sake.
Report the estimate beside the interval so the reader can see both the observed centre and its uncertainty.
What this chapter covers
- 01
Point estimates and margins of error
- 02
Z intervals for a population mean
- 03
T intervals with unknown population variance
- 04
Confidence intervals for a proportion
- 05
Repeated-sampling interpretation
- 06
Sample-size planning for means and proportions
Worked example · free
Estimate a mean service time
- 1Identify the parameter as the population mean completion time in minutes.
- 1Compute standard error as 10 divided by 5, giving 2 minutes.
- 1Compute margin of error as 2.064 times 2, giving 4.128 minutes.
- 1Form endpoints 37.872 and 46.128 minutes.
- 1Interpret the plausible range under the t-procedure conditions and sampling design.
Key terms
- Point estimate
- A single sample statistic used to estimate an unknown population parameter.
- Critical value
- A distribution cutoff chosen to produce the stated confidence level under the method.
- Margin of error
- Critical value times standard error, placed on each side of an estimate.
- Degrees of freedom
- The amount of independent information used to select a Student's t distribution.
- Confidence procedure
- A repeated-sampling method designed to capture the fixed parameter at a stated long-run rate.
Confidence Intervals and Sample Size FAQ
What does ninety-five per cent confidence mean?
It means the interval-producing procedure captures the fixed population parameter in about ninety-five per cent of comparable repeated samples under its assumptions. After one interval is calculated, its endpoints are fixed rather than carrying a new probability about the parameter.
Why is a t interval wider than a z interval?
When population standard deviation is unknown, it is estimated from the sample. The t distribution has heavier tails to represent that extra uncertainty, particularly with small degrees of freedom. The difference shrinks as sample size grows.
Why must planned sample size be rounded up?
The algebra gives the minimum continuous value needed to meet a margin requirement. Rounding down leaves too little information and produces a margin larger than promised. Fieldwork may also need extra invitations to obtain the required completed observations.
Exam move
Memorise interval families by their standard error rather than as unrelated formulas. For each problem, name the parameter, identify known or estimated variability, and predict how confidence, variability and sample size affect width. Practise interpreting the procedure without assigning a probability to a fixed parameter. Recalculate sample size for several margins and observe the inverse-square relationship.
Keep design bias separate from formula uncertainty: a narrow interval from a convenience sample does not become representative. Finish every answer with the population and unit. Build one interval from scratch by writing estimate, critical value, standard error, margin and endpoints on separate lines. Then change only the confidence level and predict the width direction before calculating.
Repeat by changing only n and only variability. For mean intervals, explain why Student's t is used when σ is estimated and record the degrees of freedom. For proportions, check that expected successes and failures support the approximation and keep proportions on either decimal or percentage-point scale throughout.
Practise the repeated-sampling interpretation aloud until it no longer sounds like a probability about the fixed parameter. In planning questions, round the calculated n upward and distinguish completed responses from invitations sent. Tie the requested margin to a business tolerance, because precision has cost only when the decision benefits from it.
Review the sampling frame beside the interval so formula precision never hides coverage or response bias.
Working through Confidence Intervals and Sample Size in BUSS1020? Sia is AskSia’s AI Statistics tutor — ask any BUSS1020 Confidence Intervals and Sample Size question and get a clear, step-by-step explanation grounded in how BUSS1020 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.