Spreadsheet software will run =AVERAGE() or =SUM() on any column filled with digits—whether those digits represent billed dollars, calendar date serials (2026-10-01 = 46296), 1–5 customer ratings, or arbitrary category codes.
EVD-01 carry-forward: Before classifying any column, confirm what one row represents (1 row per completed response, 0 blank cells, unique response_id) and make sure missing ratings stay blank (""), never 0.
| Scale | What Numbers Preserve | Valid Summaries | What Breaks | Columns in Feedback |
|---|---|---|---|---|
| Nominal | Identity and category membership only (no rank order) | Counts, `%` shares, modal category (`MODE` / `MODE.SNGL`) | `AVERAGE`, `SUM`, any magnitude ranking | `response_id`, `branch`, `postal_district`, `outcome_code` |
| Ordinal | Order / rank from low to high, with no proven equal distance between steps | Counts, `Bottom-2-Box` / `Top-2-Box` shares, modal category, odd-n `MEDIAN` (preserves middle category), equal-step-disclosed mean | `SUM`, unconditional mean without shares/assumption, assuming even-n midpoint or `PERCENTILE` ties survive nonlinear recodes | `rating_1_to_5` |
| Interval | Equal step size between units, without a true zero | `MIN`, `MAX`, span (`MAX - MIN`), `AVERAGE` (mean date or mean difference) | `SUM` (`694,470` date serials), ratio claims ("twice as warm" or "2x recent") | `response_date` |
| Ratio | Equal step size plus a true zero (`$0.00` = none) | `SUM` (`$1,800.00`), `AVERAGE` (`$120.00`), `MEDIAN` (`$120.00`), ratios ("2x higher") | None from scale rules alone (still inspect skew) | `invoice_usd` |
Worked Example (Level 1 — Modelled): Multi-Branch Customer Feedback Export (2026-10-01 to 2026-10-05)
Inspect all 15 synthetic post-service feedback responses (RSP-01–RSP-15) across North, South, and West. One row is one response (response_id is unique; zero blank cells).
outcome_code(unordered nominal categories1–4):1= Completed in-store (6rows,40.0%),2= Completed by mobile van (5rows,33.3%),3= Referred to partner shop (2rows,13.3%),4= Parts backorder cancelled with$0.00billed (2rows,13.3%).postal_district: Fictional 3-digit zones (410, 420, 435, 455, 480).rating_1_to_5: Integer1(Very Dissatisfied) to5(Very Satisfied). Blank = missing; never code missing as0.invoice_usd: Billed ticket amount in US dollars ($0.00= zero dollars billed).
| response_id | branch | response_date | postal_district | outcome_code | rating_1_to_5 | invoice_usd |
|---|---|---|---|---|---|---|
| RSP-01 | North | 2026-10-01 | 410 | 1 | 5 | 200.00 |
| RSP-02 | North | 2026-10-02 | 420 | 4 | 1 | 0.00 |
| RSP-03 | North | 2026-10-03 | 435 | 2 | 4 | 180.00 |
| RSP-04 | North | 2026-10-04 | 455 | 1 | 5 | 220.00 |
| RSP-05 | North | 2026-10-05 | 480 | 4 | 1 | 0.00 |
| RSP-06 | South | 2026-10-01 | 410 | 1 | 3 | 110.00 |
| RSP-07 | South | 2026-10-02 | 420 | 2 | 4 | 120.00 |
| RSP-08 | South | 2026-10-03 | 435 | 3 | 3 | 100.00 |
| RSP-09 | South | 2026-10-04 | 455 | 1 | 3 | 110.00 |
| RSP-10 | South | 2026-10-05 | 480 | 3 | 4 | 110.00 |
| RSP-11 | West | 2026-10-01 | 410 | 1 | 5 | 130.00 |
| RSP-12 | West | 2026-10-02 | 420 | 2 | 4 | 120.00 |
| RSP-13 | West | 2026-10-03 | 435 | 1 | 5 | 140.00 |
| RSP-14 | West | 2026-10-04 | 455 | 2 | 4 | 120.00 |
| RSP-15 | West | 2026-10-05 | 480 | 2 | 5 | 140.00 |
Compare the branch rollups below. Notice how North and South flip positions between the mean and median on these 5-response batches (providing investigation leads rather than a permanent branch ranking or proven cause), and note how entering 0 for 2 missing ratings drags the overall 56 ÷ 15 = 3.73 rating mean down to 56 ÷ 17 = 3.29 (-0.44 points).
| branch | responses | bottom_2_box (1–2) | top_2_box (4–5) | rating_mean | rating_median | mean_rank | median_rank | scale_guidance |
|---|---|---|---|---|---|---|---|---|
| North | 5 | 40.00% (2/5) | 60.00% (3/5) | 3.20 | 4.00 | 3 | 2 | RANK FLIP: REPORT MEDIAN & SHARES |
| South | 5 | 0.00% (0/5) | 40.00% (2/5) | 3.40 | 3.00 | 2 | 3 | RANK FLIP: REPORT MEDIAN & SHARES |
| West | 5 | 0.00% (0/5) | 100.00% (5/5) | 4.60 | 5.00 | 1 | 1 | ALIGNED: MEAN OK IF EQUAL-STEP STATED |
| All_Branches | 15 | 13.33% (2/15) | 66.67% (10/15) | 3.73 | 4.00 | ALL BRANCHES | ALL BRANCHES | BRANCH RANK FLIP PRESENT (2 OF 3) |
Match each column from the `Feedback` export to its measurement level and legitimate summary rule.
In `Feedback`, `=AVERAGE(E2:E16)` on `outcome_code` returns `2.00`. Why is reporting `2.00` illegitimate here?
Branch Mean Check — North's 5 ratings (5, 1, 4, 5, 1; sum = 16) give a sample rating mean of .
Hint: Divide 16 by 5 (3.20), which puts North behind South's 3.40 (17 ÷ 5) on mean rank.
Branch Median Check — Sorted from low to high (`1, 1, 4, 5, 5`), North's odd-count median (3rd of 5) is .
Hint: The middle observed rating in `[1, 1, 4, 5, 5]` is 4 (putting North ahead of South's 3.00 median).
Missing-Value Trap — Coding 2 unreturned surveys as `0` instead of leaving them blank turns `56 ÷ 15 = 3.73` into `56 ÷ 17` = .
Hint: Divide 56 by 17 and round to two decimals (`3.29`, a `-0.44`-point distortion from an impossible `0` rating).
Supported Practice (Level 2): 8-Variable Dashboard Summary Audit & Boundary Check
Audit each proposed dashboard summary below (variables 1–5 come from Feedback; 6–8 are additional site operational fields). Remember Errorless Craft: accept sound null items without inventing objections, reject scale violations by naming the exact property violated, and treat the 1–5 rating mean as a calibrated borderline case.
| # | Variable | Definition / Value Range | Proposed Dashboard Summary |
|---|---|---|---|
| 1 | postal_district | `410, 420, 435, 455, 480` (service zone code) | "Mean customer zone is 440.0." |
| 2 | outcome_code | `1`=In-store, `2`=Mobile van, `3`=Partner shop, `4`=Backorder cancelled | "Average outcome code is 2.00." |
| 3 | rating_1_to_5 | Integer `1` (Very Dissatisfied) to `5` (Very Satisfied) | "Overall mean customer rating is 3.73." |
| 4 | invoice_usd | `$0.00` to `$220.00` billed ticket amount | "Total billed revenue is $1,800.00; mean ticket is $120.00." |
| 5 | response_date | `2026-10-01` to `2026-10-05` | "Responses span 2026-10-01 through 2026-10-05 (4 days)." |
| 6 | wait_minutes | Minutes waited before service (`0` to `45`; `0` = immediate) | "Median wait is 12 min; mean wait is 14.5 min." |
| 7 | membership_tier_code | `1`=Bronze, `2`=Silver, `3`=Gold (`60%` Gold, `25%` Silver, `15%` Bronze) | "Majority tier is Gold (60%), with 85% in Silver or Gold." |
| 8 | workshop_temp_celsius | Ambient bay temperature in °C (`12°C` morning vs. `24°C` afternoon) | "Afternoon bay temperature (24°C) is twice as warm as morning (12°C)." |
Audit Row 1 (`postal_district`): "Mean customer zone is 440.0." What is the accurate classification and verdict?
Audit Row 2 (`outcome_code`): "Average outcome code is 2.00." Which audit response is correct?
Audit Row 3 (`rating_1_to_5`): "Overall mean customer rating is 3.73." How does EVD-02 classify this calibrated borderline item?
Audit Row 4 (`invoice_usd`): "Total billed revenue is $1,800.00; mean ticket is $120.00." What is the correct audit verdict?
Audit Row 5 (`response_date`): "Responses span 2026-10-01 through 2026-10-05 (4 days)." How should you rule?
Audit Row 6 (`wait_minutes`): "Median wait is 12 min; mean wait is 14.5 min." What is the verdict?
Audit Row 7 (`membership_tier_code`, `1`=Bronze, `2`=Silver, `3`=Gold): "Majority tier is Gold (60%), with 85% in Silver or Gold." What is the verdict?
Audit Row 8 (`workshop_temp_celsius`): "Afternoon bay temperature (24°C) is twice as warm as morning (12°C)." Why does this fail?
Boundary Check: How should you brief operations leaders on `North` (`3.20` mean / `4.00` median) vs. `South` (`3.40` mean / `3.00` median) from these 15 rows?
Executive English Briefing Prompt (CEFR B2 → C1 · Discord Practice)
Write a 3-sentence executive update for an operations review meeting explaining the North vs. South rank flip using the exact numbers (West 4.60 mean / 5.00 median, 100% 4–5; North 3.20 mean / 4.00 median, 60% of responses 4–5, 40% 1★ cancelled backorders at $0.00; South 3.40 mean / 3.00 median, 60% 3★).
- Don't say: "Ordinal rating means fail strict interval assumptions across Likert scales, and
Northdrops from Rank 2 to Rank 3 on an uncalibrated parametric mean whileoutcome_codeis nominal." - Do say: "Each branch only has 5 responses in this batch, so treat these numbers as leads to check rather than a branch ranking. West leads on both average (4.60) and middle score (5.00), with 100% of its responses at 4 or 5. North averages lower than South (3.20 vs. 3.40) because two 1-star responses pull its mean down, even though North's middle score is 4.00 and 60% of North's responses are 4 or 5—and those two 1-star rows were cancelled backorders at $0.00 billed, so the first follow-up is checking parts availability at North and visit notes behind South's 3-star scores."
- Cohort route: Post your 3-sentence briefing draft in
#track-1-evidenceor#practice-and-troubleshooton Discord for peer and instructor feedback.