Are three of your regions really scoring the same?
Paste the ratings for each region and get back whether the regions really differ. Four regions, twelve star ratings each, on a one-to-five scale. Three regions share a median of 4 and the test still separates them, because a median throws away everything a rank test keeps.
Words on this sheet
- Median: The middle value: half the readings sit above it and half below.
- Mean: The average: everything added up and divided by how many there are.
Operations Starter Statistics free
After you install, this is the model to open.
Do the Regions Rate Us Differently?
- In your spreadsheet, click the Sortia icon in the strip of icons down the right-hand edge. No strip? Click the arrow at the bottom-right to open it. You can also use Extensions, then Sortia, then Open Sortia.
- Click Start from a template and put that name in the search box.
- Pick the card with that name and click Load this template. It arrives on a new tab with real numbers already in it.
The answer
- The verdict
- p = 0.00095 H 16.37 on 3 degrees of freedom
- East
- 12.6 mean rank of 48 responses, the obvious problem
- South
- 22.8 mean rank; its median is still 4, the quiet one
- West and North
- 32.3 / 30.3 mean ranks at the top
Four regions, twelve customer ratings each, one to five stars. The summary block at the side of the sheet says three of the four regions have a median of 4, so a dashboard built on medians shows three identical regions and one weak one. Click Run: H comes back 16.37 on 3 degrees of freedom with a p-value of 0.00095, and the mean-rank column separates all four, West at 32.29, North at 30.25, South at 22.83 and East at 12.63.
South and North both have a median of 4 and sit seven and a half rank places apart, because a median only reports which value is in the middle while a rank test uses every response. The share-rating-5 row shows exactly what the median discarded: five of North's twelve customers and five of West's gave five stars, South managed one in twelve, and East gave none at all.
East is the obvious problem and South is the quiet one, and the median hid the second entirely. Star ratings are ordered labels rather than distances, which is why this is the right family of test. Nobody can defend the claim that the gap between four and five stars is the same quantity as the gap between one and two, and an analysis of variance on this data would assume exactly that.
Read the result for what it is: the test says the four regions are not all the same and does not say which pairs differ, so follow it with Mann-Whitney on North against South before anyone writes a plan. Twelve responses a region is thin, and rerunning on a month rather than a week will move the mean ranks around less between runs. What no test can fix is who chose to leave a review at all: the people who had an ordinary delivery mostly said nothing, and that selection is bigger than anything on this sheet.
To use your own data, put one region per column with its name in the top cell of that column and widen the range to your rows.
The model
It arrives on a tab called Template: Do the Regions Rate Us Differently, carrying these columns:
- North (review stars 1-5)
- South (review stars 1-5)
- East (review stars 1-5)
- West (review stars 1-5)
- West (stars, 1-5)
with the model computed beside the data:
| Median | 4 |
| Mean | 4.167 |
| Share rating 5 | 0.4167 |
Once it is in your sheet
- The model arrives with real numbers in it and runs as it stands, so you can press the button first and understand it second.
- Change the numbers to yours. The sheet marks which cells are inputs and which hold formulas, and most labels carry a note explaining the row.
- Press the run button at the bottom of the panel. It is labeled for the tool you are in, and the result lands on its own tab, with a written reading of it beside the figures.
Never used Google Sheets? Start here goes the whole way, in seven steps, and assumes nothing.
Next question
- The season looks fine on pace. Is it really?Will the Season Fill?
- Should weekends be staffed like any other day?Weekday and Weekend Are Not the Same Business
- What is the standing recipe costing you a tonne?The Cheapest Blend That Meets the Spec
- Will the changeover fit inside the weekend?How Much Buffer Does the Changeover Need?
- Same average fill. So why is one supplier a risk?Which Supplier Is More Consistent?
- Two crews average the same. Is one of them better?Four Shifts, Ranked Output, One Question
Every model like this one, and the method behind them: Statistics in Google Sheets.
The method behind this one, worked end to end: ANOVA in Google Sheets, and Kruskal-Wallis beside it.