MGMT7250 Chap.5 Samples, Intervals and Trustworthiness
Samples, Intervals and Trustworthiness
A sample statistic becomes decision evidence only after the target population, sampling frame, selection route, measure and denominator are clear. This chapter computes rates, reads confidence intervals against action thresholds and separates sampling precision from broader trustworthiness.
It shows why a large convenience sample can estimate the wrong target precisely, why the confidence belongs to the repeated method, and why every quantitative conclusion should name residual uncertainty with a direction.
Separate the target population, operational frame and realised sample. Exclusion can occur because units never enter the frame, do not respond or are removed during processing.
The critical question is whether inclusion relates to the outcome. A survey of active application users may estimate their preferences precisely while excluding the customers most likely to resist an application-only service.
Always report numerator and denominator. Rates can move because events change or because the opportunity count changes, and raw counts conceal that distinction.
A confidence interval describes repeated-sampling variation under the method; it does not repair selection, coverage, classification or under-reporting. Those uncertainties require separate appraisal and a directional caveat.
Read the entire interval against a decision threshold chosen before seeing the result. When an interval crosses tolerance, the point estimate alone cannot establish acceptability.
When every plausible value supports the same reversible action, a wider interval may still be decision-sufficient. This links statistical precision to managerial consequences rather than visual narrowness. State the smallest difference that matters, the action at each end of the range and the observation that would most reduce remaining uncertainty.
Quantitative reasoning is complete only when the calculation, design boundary and decision consequence appear together.
What this chapter covers
- 01
Target population and unit
- 02
Sampling frame and selection
- 03
Rates and denominators
- 04
Confidence interval interpretation
- 05
Decision thresholds
- 06
Precision versus trustworthiness
Worked example · free
Interpret an interval that crosses tolerance
- 1Compute 34 divided by 800 as 4.25%.
- 1Place the estimate below the 5.0% threshold.
- 2Note that the interval crosses the threshold and prevents an acceptable-performance conclusion.
- 1Name coverage or measurement uncertainty separately.
Key terms
- Target Population
- The Target Population is the full set of units to which a decision or inference is meant to apply.
- Sampling Frame
- The Sampling Frame is the set of units that could actually be selected for the sample.
- Confidence Interval
- A Confidence Interval is a range generated by a procedure with a stated long-run coverage rate.
Samples, Intervals and Trustworthiness FAQ
Why is a point estimate not enough?
The point hides sampling uncertainty and can sit on the opposite side of a decision threshold from plausible population values. An interval shows the range the method supports under its assumptions.
Does a large sample guarantee trustworthy evidence?
No. Size can improve precision inside the observed frame, but it cannot include excluded groups, repair a shifting definition or remove systematic under-reporting. Trustworthiness must be appraised before celebrating a narrow interval.
How should I describe a confidence interval?
Report the range and explain that the method has the stated long-run coverage under its assumptions. Do not assign a probability to the fixed population parameter, and state frame or measurement limitations separately.
Assessment move
For every number, write a seven-part audit: estimand, unit, population, frame, selection, measure and threshold. Recompute the rate from numerator and denominator, sketch the interval, then read the lower and upper ends aloud against the action rule. Finish with one directional caveat that the interval cannot capture.
Draw three nested boxes for target population, frame and sample, then annotate who is lost at each boundary and the likely direction of bias. Compare two cases with identical sample sizes but different coverage to explain why precision is not trustworthiness. Practise translating interval language without assigning probability to the fixed parameter.
Finally, alter the denominator, threshold and decision reversibility one at a time and explain how the recommendation changes even when the point estimate remains constant. Reconstruct three published percentages from their numerators and denominators and flag any missing exposure information. For each, write separate sampling, coverage and measurement uncertainties.
Sketch a threshold before plotting the interval so the action rule cannot move after seeing the estimate. Compare the decision at the lower endpoint, centre and upper endpoint, then identify which new sample would narrow the most consequential uncertainty. End by drafting a quantitative conclusion containing the population, rate, interval, threshold, residual bias direction and operational implication in that order.
Working through Samples, Intervals and Trustworthiness in MGMT7250? Sia is AskSia’s AI Management tutor — ask any MGMT7250 Samples, Intervals and Trustworthiness question and get a clear, step-by-step explanation grounded in how MGMT7250 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.