Validation of a sampling plan to generate food composition data

BASSETT M.N; GIMENEZ M.A; MARCOLERI M.E; SAMMAN N.C; LOBO M.O

FOOD CHEMISTRY

ELSEVIER SCI LTD

Lugar: Amsterdam; Año: 2016 vol. 193 p. 141 - 147

0308-8146

Abstract: A methodology to develop systematic plans for food sampling was proposed (FAO project TCP/RLA/3107). Long life whole and skimmed milk and sunflower oil were selected to validate the methodology in Argentina. Fatty acid profile in both foods and proximal composition and calcium content in milk were determined with AOAC methods. The number of samples (n) was calculated applying Cochran´s formula with coefficients of variation ≤ 12% and maximum permissible estimate error (r) ≤ 5% for calcium content in whole and skim milk and unsaturated fatty acids in oil. The n were 9, 11 and 21 for long life whole and skimmed milk and sunflower oil respectively. Sample units were randomly collected from production sites and sent to labs for analysis. Calculated r with experimental data was ≤10% indicating high accuracy and reliability. The proposed methodology is an adequate and useful tool for developing sampling plans for food composition analysis purposes.