Results table
The table of mean ± SD per attribute and sample, with an ANOVA column and Tukey letters, that accompanies the spider chart.
What it is
The results table is the numeric detail that accompanies the spider chart in descriptive and hedonic tastings. Where the spider gives you the visual read of the profile, the table gives you the exact values and the significance test, attribute by attribute.
How to get there
- URL:
https://app.areacacao.com/sensory-analysis/evaluations/<id> - How to reach it: open the tasting under Evaluations and scroll down to the Results section, below the spider chart and the sensory sample map (PCA).
Pro plan or above, or a tasting pack with credit left.
How it is organized
- Above the table, a summary: number of panelists, samples, and attributes.
- Each row is a protocol attribute.
- Each column is a sample, identified by its blind code.
- In each cell: the mean ± standard deviation (SD) of the panel for that attribute and that sample.
- An ANOVA (p value) column per attribute, indicating whether the attribute varies across samples.
Tukey letters
A superscript letter (ᵃ, ᵇ, ᶜ) may appear next to each mean. These are the Tukey groups: samples that share a letter do not differ significantly from one another. It is the quick way to see which samples are indistinguishable to your panel and which ones stand out.
ANOVA column
| Marker | Meaning |
|---|---|
| Difference (green) | There is a statistically significant difference between samples for that attribute. |
| No difference (gray) | No significant difference is detected. |
| n/a | Not enough data to compute the test. |
Below the marker, each cell in the column shows the p value (for example, p = 0.012, or p < 0.001) and the statistic with its degrees of freedom (for example, F(2; 18) = 5.40). The test is only computed for an attribute with at least 3 samples scored by at least 3 panelists each; otherwise you'll see n/a.
A note at the foot of the table reads:
Mean ± standard deviation per sample. Letters (ᵃ ᵇ ᶜ) indicate Tukey groups: samples sharing a letter do not differ significantly (α = 0.05).
Below it, another note states which model was used: a two-way ANOVA with the panelist as a block when at least 3 panelists scored every sample (each judge's bias is removed from the error term), or a one-way ANOVA when there aren't enough. If any panelist with incomplete data was left out of the test, the note says how many.
How to read it
- First locate the attributes marked Difference (green): those are the ones that set your samples apart. The No difference ones are the common ground.
- Use the Tukey letters to cluster: samples with the same letter on an attribute are interchangeable for that attribute according to your panel.
- Always check the SD: a high mean with a high SD means the panel does not agree, and the conclusion should be qualified.
The table shows aggregates (mean ± SD per attribute and sample), not the raw responses of each panelist one by one, and it can't be exported to CSV or Excel from the results screen. You review the analysis inside the tasting itself.