The task and original case
Compare a survey summary based on responses with one based on declared population shares. The difference is a useful research portfolio exercise because it forces you to name the population you intend to describe. Weighting a convenience sample does not prove that its respondents represent everyone.
| Case element | Value or requirement |
|---|---|
| Inputs | Invented populations A 900/B 100;20 responses in each; positive A 10/B 18. |
| Build | A weighting workbook and a clear limit on what a convenience sample establishes. |
This is an original invented exercise. It is independent practice, with no employer assessment or open-vacancy claim. Download the inputs, attempt your own result, then open the separate answer key.
Data and worked inputs
Build a result another person can follow
Create one row per group with population, responses and positive responses. A has population 900, 20 responses and 10 positives; B has population 100, 20 responses and 18 positives. First calculate the positive share within each group. Separately calculate each group's share of the total invented population.
Multiply within-group positive share by population share, then add those contributions. Keep that weighted calculation beside the unweighted response calculation. If you use a PivotTable to summarize inputs, verify that numeric fields use the intended Sum or Count setting; Microsoft's documentation explains that choice. The weighting method in this exercise is an authored arithmetic model, not a statistical claim from the spreadsheet documentation.
Sources[1]
Check the worked result
A's positive share is 10 ÷ 20 = 50%; B's is 18 ÷ 20 = 90%. Population shares are 90% and 10%. The weighted summary is 0.9 × 0.5 + 0.1 × 0.9 = 0.54, or 54%. The response summary is 28 ÷ 40 = 70%.
Both calculations are internally valid for their declared weighting rules. Their 16-percentage-point difference comes from which groups receive influence. The weighted 54% still relies on an unsupported assumption that respondents within each group resemble that group's population.
Make the failure visible
Change B's population to zero while keeping its responses. Flag the inconsistency and decide how the analysis population should be repaired; do not simply hide the group. Test a group with no responses too. Its positive share is unknown, so a positive population weight cannot manufacture its missing outcome. Record how excluding it changes the scope of the result.
Revise it and ask someone to reproduce it
For revision two, increase A's positive count from 10 to 12. Predict the weighted contribution before recalculating: A's share becomes 60%, giving an overall 63%. Ask a peer to identify the representativeness assumption in your caption. Rewrite the caption if they mistake weighting for evidence of a representative survey.
- Group shares, population weights and weighted contributions reconcile.
- The report identifies the convenience-sample limitation.
- Unknown group outcomes remain visible rather than becoming zero.
Describe the work accurately
After completing your own version, you could adapt this project bullet: “Compared response-weighted and population-weighted summaries in an invented survey; documented the sampling assumption and missing-group limits.”
Use Projects or Independent practice as the heading. Keep the input files, result and revision together. Describe only your own completed work; the invented numbers are model outputs, not employer results or measured business impact.