Quality assurance, sensory evaluation and continuous improvement
Scope
This domain tests quality assurance, sensory evaluation and continuous improvement: specifications, sampling, statistical process control, discrimination and preference tests, root-cause analysis and CAPA.
Core rules
- A specification states measurable attributes: weight, temperature, pH, colour, texture, microbiological limits.
- SPC uses control charts to separate common-cause and special-cause variation. Control limits are not specification limits.
- Sensory tests: triangle and duo-trio test discrimination; paired comparison and hedonic tests measure preference or acceptance.
- Panels need training, controlled lighting, coded samples and randomised order.
- Continuous improvement uses PDCA, root-cause analysis (5 Whys, fishbone) and verified corrective action.
Worked example
A triangle test with 30 panelists gives 16 correct identifications. Chance probability is 1/3, so expected correct = 10. The critical value at 5% significance for n=30 is about 15. Since 16 exceeds it, the difference is significant: the new recipe is detectably different. That does not prove consumers prefer it; run a hedonic test.
Traps
- Confusing discrimination with preference.
- Using specification limits as control limits.
- Serving samples in the same order to all panelists.
- Closing a CAPA without verifying effectiveness.