FOOD9102 Chap.2 Controlling Bias in Sensory Test Design
Controlling Bias in Sensory Test Design
Panellists are human, so every sensory test is exposed to bias. This chapter defines bias as anything other than the tested factor that changes perception, works through associations, sensory fatigue, the halo effect and order effects, and shows the design controls the Week 1 lecture teaches: identical samples, separated booths, balanced serving orders, random three-digit codes, sensible timing and a standard serving tray.
A deliberately flawed tasting plan is diagnosed and repaired step by step.
The chapter is also the foundation for the laboratory reports, which ask you to comment on whether the serving order you were given followed correct experimental design and to explain the biases you met. Every control here appears again in the brief.
What this chapter covers
- 01
What bias is and why panellists bring it
- 02
Associations, fatigue and the halo effect
- 03
Order bias and balanced serving orders
- 04
Panel sizes that fit two-sample and three-sample designs
- 05
Random three-digit codes instead of letters
- 06
Timing, palate cleansing and pressure from onlookers
- 07
Laying out a serving tray
Worked example · free
Plan the serving orders for a three-sample hedonic test
- 1Three samples give 3 × 2 × 1 = 6 serving orders: ABC, ACB, BAC, BCA, CAB and CBA.
- 1Each order must be used equally, so the usable panel is a multiple of six: 24 results, four per order.
- 1The 25th result breaks the balance and is discarded, which is why a run sheet records every panellist's order.
- 1Label each muffin with a random three-digit code checked for patterns, never with letters or 1, 2, 3.
Key terms
- Halo effect
- The halo effect is the spill-over of a strong overall impression onto ratings of individual attributes.
- Sensory fatigue
- Sensory fatigue is the dulling of perception after tasting or smelling too many samples, which lowers liking scores.
- Order bias
- Order bias is the tendency to rate a sample differently because of its position in the tasting sequence, often favouring the first.
- Serving order
- A serving order is the sequence in which one panellist receives the samples, balanced across the panel.
- Run sheet
- A run sheet is the record of which panellist received which coded samples in which order.
- Expectoration
- Expectoration is spitting out a sample after tasting, used when there are many samples or strong flavours.
- Palate cleanser
- A palate cleanser is water or a plain cracker used between samples to reduce carry-over from one sample to the next.
Controlling Bias in Sensory Test Design FAQ
Why should samples carry three-digit codes rather than letters?
Letters and single digits carry built-in rankings and associations: people read A as better than B, or link a letter to a fruit. Randomised three-digit codes carry no meaning, provided you check the set for accidental patterns such as mirrored numbers or one code that stands out from the rest.
What do I do if my panel is not a multiple of the number of orders?
Keep every order equally represented. The lecture's rule for a three-sample test that ends with 37 panellists is to discard the last odd result, which leaves 36 and six per order. It is better still to recruit to a suitable multiple in the first place.
Is colour always a bias in a sensory test?
No. If colour is one of the attributes consumers respond to, it is information you want. It becomes bias when the test is about something else, such as sweetness, and a colour difference between samples nudges the panellists towards an answer.
What time of day is best for a tasting?
The lecture warns against first thing in the morning, the period straight after lunch and the end of the working day, because hunger and mood move ratings. Inside a company, avoid the weeks before performance reviews, when staff may score to please.
How many samples should a panellist taste at once?
Few enough to avoid sensory fatigue, which tires the senses and lowers liking as more samples are tasted. The lecture does not fix a number, but it ties the limit to fatigue, and it suggests palate cleansing with water or plain crackers and expectoration when there are many samples or strong flavours.
Assessment move
Build a two-column table of biases: in one column how the bias changes what the panellist perceives, in the other the design control that prevents it. Then practise on short scenarios, including your own lab plans, by finding at least three biases and fixing each. Rehearse the arithmetic of serving orders for two and three samples until you can size a panel in seconds.
In the laboratory, keep your own short log of anything that differed between samples or panellists, since it becomes the evidence for your discussion of bias. Before Quiz 1, practise turning three listed biases into explained ones, because the feedback on that quiz singled out lists without explanation.
Working through Controlling Bias in Sensory Test Design in FOOD9102? Sia is AskSia’s AI Science tutor — ask any FOOD9102 Controlling Bias in Sensory Test Design 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.
