Selected Descriptive Analysis Case Studies
Shelf Life of Brownies
Background and Objective
A client had developed six new snack products. All were brownie based with variations in toppings. Four packaging variations were utilized. The objective was to determine which packaging best maintained the sensory characteristics of the products. The goal was a 90-day post production shelf life.
Test Design
Descriptive analysis ballots were developed, and the initial test was done at two weeks after production. Products were stored at 70F/50RH and sensory attributes tested at predetermined intervals.
Results
Mold developed in all packaging samples between the fourth and seventh week time points; therefore, subsequent testing was terminated. Excluding the samples that displayed the mold, the samples stored in Package B and Package D were most like the initial test brownies. Samples in packages A and C indicated a stronger likelihood for developing off flavors than the other two samples.
Conclusions
Package B or Package D can be utilized for a short shelf life. Further package development is needed to reach the target 90 days.
Product Matching of Shellfish Patties
Background and Objective
A restaurant chain was seeking to qualify an alternate vendor for their shellfish patties. The objective was to determine whether the alternate product was a match to the product currently being served or fell within the range established by multiple control samples.
Test Design
Two control and one test sample were submitted for evaluation by the descriptive panel. After developing a ballot listing flavor and texture attributes of the products, each sample was rated, attribute by attribute.
Results
While the two control samples were nearly identical, the test sample was not a match to the controls. The test sample displayed more crab flavor and less grain or starch flavor. It had pieces of cooked onion not found in the controls, and the texture was firmer and more cohesive and fibrous.
Conclusions
Further development is needed by the proposed alternate vendor to produce a product matching the target, control profile.
Insurance Claim for Tainted Butter
Background and History
A client insured a warehouse containing 12,000 pallets of frozen and refrigerated butter, both salted and unsalted, in bulk and in sticks. After an increasing number of customer complaints, the butter manufacturer wanted to declare the product a loss due to taint. On inspection of the warehouse, it was noted that fruit concentrates used in mixed drinks were stored in the same warehouse with the butter; as a result, a fruity aroma was perceived in the immediate area of the butter.
Test Design
Using grids of the building and pallet layout, a statistician calculated the total square root of all the pallets and pulled from them random samples at the top, middle and bottom; this data provided the design for sampling. The high and low end of the data were anchored by an unstored control sample of butter and an induced sample of fruit flavoring.
A total of 495 control and test butter samples were tested by a pool of 22 trained descriptive panelists. The five attributes on the ballot were dairy/butter, diacetyl, fruity, salt and oxidized. The test was designed so ten panelists tested sets of ten samples each for a total of 4,950 evaluations.
Results
The control sample containing induced taint had the highest fruity intensity score. The control sample that was not stored in the warehouse had no fruity flavor. All of the test samples, except one bulk sample stored in a remote freezer in the warehouse, were rated significantly higher than the Control group in fruity.
Conclusions
The data demonstrated that a fruity flavor taint was present. The butter was declared a loss.
Masking Diaper Pail Aroma
Background
The client had developed a diaper containment system (Prototypes #1 and #2) that utilized activated charcoal and a small fan. The fan circulated the air while the activated charcoal removed odors generated by the soiled diapers.
Test Design
To test how well the new system contained diaper odors, a small, controlled space simulating a home bathroom was needed. 21st Sensory contracted an engineering company to build five, 5-foot square cubes of wood with apertures through which a person could sniff the air inside. Soiled diapers were then collected from local daycare centers, sorted and evenly distributed among four of the diaper containments systems. The fifth unit was left empty to provide a blank control.
A test was designed that used twelve trained descriptive panelists to sniff the air inside the cube. The air was tested in 24 hour intervals. Cubes were coded, randomized and moved between replications so panelists could not develop knowledge, expectation or bias about a particular cube's aroma.
After completing the aroma evaluation, the data were statistically analyzed to determine if there were significant differences in the odor intensities between the five cubes.
Results and Discussion
As predicted, the control cube was rated as having the least odor. It was characterized as having a slight 'new wood' aroma. Competitor #1 had an added perfumed, 'baby powder' scent. Initially, the added fragrance was dominant. However, on the second and third days, the fecal and urine smells became stronger than the added scent.
Competitor #2 performed the best in concealing the odor.
Prototype #2 performed as the second best amongst all samples, second only to the Competitor #1. The study data demonstrated that Prototype #2 performed better than Prototype #1 in reducing fecal and urine odors. This efficacy was attributed to Prototype #2s increased power (four 'D' batteries in Prototype #2 versus three 'D' batteries in Prototype #1) and the resulting increased volume of airflow.
LSD Means Separation Analysis of Aroma Intensity Scores
|
Control |
Competitor #1 |
Competitor #2 |
Prototype #1 |
Prototype #2 |
||||||
|
Day One |
3.2 |
c |
6.6 |
a |
5.0 |
b |
5.1 |
b |
5.5 |
b |
|
Day Two |
3.1 |
e |
7.7 |
a |
4.6 |
d |
6.9 |
b |
5.5 |
c |
|
Day Three |
3.7 |
e |
8.0 |
a |
4.8 |
d |
7.4 |
b |
5.3 |
c |
*Samples in the same row with different letters are significantly different at 0.5.