Glucose and fructose as sugar sources were added to raisin concentrate called Pekmez in Turkey at the ratio of 10-25% and 50% w/w, to simulate adulteration in order to develop a statistical regression model. Sugar composition was analyzed with standard enzymatic methods and mineral matters were determined with atomic absorption spectrophotometry. According to the results of analysis for prepared raisin concentrate with different ratios, adulterants could be determined in the limits of variation for each composition component. The minimum variations for these values were fixed and regression models were developed. These statistical models were used in authentication. Regression analysis method could be applied in quality control of raisin concentrate or in monitoring adulteration or contamination. (C) 2003 Elsevier Ltd. All rights reserved.