SOP: Two-Sample Difference Test

GIFSQ — Global Institute for Food Safety & Quality

Principle

A variation of the paired comparison test that measures the amount of difference between a test sample and a standard. Each taster is served four pairs of samples; each pair consists of an identified reference and a coded test sample. In two pairs the test sample is a duplicate of the reference (coded as a check on guessing); in the other two pairs the test sample is the test variable. The panelist is asked to judge each pair independently for the degree of difference and, additionally, the direction of the difference. The panelist must not guess: the two pairs containing a coded duplicate of the standard check whether guessing has occurred.

Apparatus required

• Individual judging booths (identical, uniformly illuminated), each fitted with drinking water, a glass, clean towels and a basin or receptacle for expectoration
• Coloured lighting (where needed to mask small colour differences between samples)
• Stainless steel, glass or china serving dishes and cups; plain serving trays; small stainless steel forks and spoons
• Prepared evaluation cards for the two-sample difference test

Sample preparation

Prepare the reference (R) and the test sample (T) in exactly the same manner. Prepare coded duplicates of the reference (R′). Make up four pairs per panelist: two pairs RR′ (reference plus coded duplicate) and two pairs RT (reference plus test sample), arranged in a randomized design. Code all test samples with three- or five-digit numbers.

Procedure

Serve each of the four pairs to the panelist; in each pair the reference sample is identified and the test sample is coded (trained panel: 6–25 judges; 4 pairs of unknown and control samples per session)

↓

For each pair independently, compare the coded sample with the reference sample

↓

Determine the degree of difference on the scale: 0 no difference, 1 very slight difference, 2 moderate difference, 3 large difference

↓

Determine the direction of the difference: S superior to standard, E equal to standard, I inferior to standard (if there is no difference, there is no degree or direction)

↓

Comment on what the difference is based on: odour, taste or both

↓

Assign numerical ratings: equal to R = 0; large difference = +3 (superior) or −3 (inferior); moderate difference = +2 or −2; slight difference = +1 or −1

Interpretation of results

The difference may be evaluated either by the t-test-by-difference or by the analysis of variance.

t-test-by-difference:

t = ΣD / √{[N·ΣD² − (ΣD)²] / (N − 1)}

• ΣD = algebraic sum of the differences between the paired scores
• ΣD² = sum of the squares of the differences
• N = number of pairs of scores
• Note: the equation layout is garbled in the available copy of the book; the form given here reproduces the book’s worked values (t = 0.71 for the first presentation and t = 0.387 for the second, both below significance at the 5% level, Table 19-11). Check: (10 × 47 − 5²)/(10 − 1) = 445/9; √ = 7.032; 5/7.032 = 0.71.

Analysis of variance: tabulate the ratings for all four pairs by all panelists; compute the sums of squares for samples, panelists and error; test the F ratios against Tables 19-6 and 19-7.

Worked illustration from the book: to test whether jam made with sugar (R) differed from jam with 20% of the sugar replaced by dextrose syrup (T), 10 panelists judged four pairs (RR′, RT) in a randomized design. By the t-test, the first presentation gave t = 0.71 (9 degrees of freedom) and the second t = 0.387 — not significant at the 5% level; combining both presentations gave t = 0.785 (19 degrees of freedom) — not significantly different. The analysis of variance confirmed the same conclusion, and further showed the panelists were consistent in their scoring. The positively correlated scores remove the effect of differences between levels of scores from the estimate of error, as in the analysis of variance.

Specimen evaluation card fields: instructions (compare each coded sample with the reference independently; determine degree and direction); table of sample code number, degree of difference, direction, comment; signature.

Precautions

• Present the pairs in a randomized design; each panelist must judge every pair independently.
• Do not let the panelist guess — the coded duplicate standards in two of the pairs penalize guessing.
• Conduct judging in individual booths so that judgements are independent; allow no communication between panel members except consultation with the panel leader on any point of doubt.
• Test in the hour after a normal meal (e.g., 10.00 a.m. to 12 noon, or 3 to 5 p.m.).
• Panel members should not smoke or chew pan or supari for at least half an hour before the test.
• Judge quickly but not hurriedly; make odour observations by sniffing before tasting.
• Allow retesting, but avoid prolonged testing; tasters may swallow or spit out samples, and should rinse between samples (or use bread or puffed rice) to remove a strong flavour.
• Keep the evaluation card simple, clearly typed or printed, with unambiguous terms and directions in the desired sequence of action.

References

Ranganna, S., Handbook of Analysis and Quality Control for Fruit and Vegetable Products, 2nd edition.