FOOD9102 Chap.8 JAR Scales and Penalty Analysis
JAR Scales and Penalty Analysis
A just-about-right scale tells a developer whether an attribute is too weak, too strong or right, and penalty analysis says which of those faults matters most. This chapter covers unipolar and bipolar JAR scales, the 75% benchmark, the mean drop calculated against the JAR group’s liking, and the penalty plot of mean drop against the proportion of panellists.
A two-attribute iced tea example is worked from raw groups to a reformulation priority.
Penalty analysis combines tools from earlier chapters: the JAR scale from Chapter 6 and the mean liking scores of Chapter 7. It is the course’s main bridge from data to a reformulation decision.
What this chapter covers
- 01
Just-about-right scales and their ideal point
- 02
Unipolar and bipolar layouts
- 03
The 75% JAR benchmark
- 04
Pairing JAR ratings with overall liking
- 05
Mean drop and proportion of panellists
- 06
Reading the penalty plot to set priorities
Worked example · free
Penalty point for a dry muffin
- 1Benchmark: 55% JAR is below the 75% benchmark, so a penalty analysis is justified.
- 1Mean drops: too dry 7.8 − 5.9 = 1.9 at proportion 0.35; too moist 7.8 − 7.1 = 0.7 at proportion 0.10.
- 1Priority: too dry is both costlier and more common, so make the muffin moister first.
Key terms
- Just about right
- Just about right is the central category of a JAR scale, meaning the attribute's intensity is ideal.
- JAR benchmark
- The JAR benchmark is the target that three quarters or more of panellists, 75%, choose the just-about-right category.
- Mean drop
- Mean drop is the JAR group's mean liking minus the mean liking of a non-JAR group.
- Penalty plot
- A penalty plot graphs each fault's mean drop against the proportion of panellists reporting it.
- Bipolar scale
- A bipolar scale runs from much too weak to much too strong with the ideal at its centre.
- Unipolar scale
- A unipolar JAR scale runs from not at all strong enough through just about right to much too strong.
JAR Scales and Penalty Analysis FAQ
Why does a JAR test need a hedonic question as well?
The JAR question alone shows how many people think an attribute is off, but not how much that costs in liking. Pairing it with an overall liking score lets you calculate a mean drop for each fault, which is what a penalty analysis ranks.
When is a penalty analysis most useful?
The lecture recommends it for entirely new products, where there is no good benchmark, and for major reformulations, especially when several attributes are assessed at once. It tells you which improvements matter most, not how to make them.
What does a point in the top right of a penalty plot mean?
It is a fault that lowers liking a lot and is reported by many panellists. Those points are where reformulation effort should go first, while points low down or far to the left can usually wait.
What does the 75% JAR benchmark mean?
It is the target the lecture sets for a well-optimised attribute: at least three quarters of panellists should rate it just about right. An attribute below that benchmark is a candidate for change, and a penalty analysis then shows whether the fault is costly enough, and common enough, to act on.
Assessment move
Rebuild a penalty analysis by hand from a small table: sort the panellists into groups for each attribute, check the groups add up to the panel, compute each mean drop against the right JAR mean, and plot the points. Then write a two-sentence recommendation that names the attribute and the direction of change. Practise the full method on a second attribute so that you see how a different JAR group changes the reference mean.
In a report, present the plot and a short table of drops and proportions together, so the reader can check every point.
Working through JAR Scales and Penalty Analysis in FOOD9102? Sia is AskSia’s AI Science tutor — ask any FOOD9102 JAR Scales and Penalty Analysis question and get a clear, step-by-step explanation grounded in how FOOD9102 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.
