Vol. 69, Nr. 9, 2004—JOURNAL OF FOOD SCIENCE S401 Published on Web 12/1/2004 © 2004 Institute of Food Technologists Further reproduction without permission is prohibited S: Sensory & Nutritive Qualities of Food Consumer Acceptability Compared with Sensory and Instrumental Measures of White Pan Bread: Sensory Shelf-life Estimation by Survival Analysis A. GÁMBARO, S. FISZMAN, A. GIMÉNEZ, P. VARELA, AND A. SALVADOR ABSTRACT: Two studies, one in Uruguay using Uruguayan white pan bread (WPB) and another in Spain using Spanish WPB, were carried out with trained judges and consumers from the country in question. The objectives were to correlate the consumer panel acceptability, trained sensory panel scores, and instrumental measure- ments of defects likely to appear during WPB storage, to compare the acceptability prediction models obtained in each country, and to apply survival analysis methodology to estimate the shelf life of the product. Logistic regression was used to predict the acceptance percentage and multiple regression was used to predict accept- ability. Sensory variables (strange odor) showed a greater prediction capacity for Uruguayan WPB acceptability and acceptance percentage. For the Spanish WPB, instrumental cohesiveness and moisture content variables showed a greater prediction capacity for the acceptance percentage, whereas sensory (cohesiveness and color) and moisture content variables showed a greater prediction capacity for acceptability. Uruguayan WPB shelf life was 15 d for a 50% probability of consumer rejection and 13 d for a 25% probability of consumer rejection, whereas for Spanish WPB it was 23 d for a 50% probability of consumer rejection and 16 d for a 25% probability of consumer rejection. Keywords: acceptability, white pan bread, shelf life, survival analysis Introduction S ensory evaluation is the key to determining the shelf life of white pan bread (WPB). As far back as 1955, Betchel defined the aging of bakery products as the decrease in consumer acceptance caused by changes in the crumb that were not the result of microbiological action. Of all the changes that take place in bread, crumb hardening has been the most widely used parameter for evaluating the degree of aging in storage. Instrumental (compression) measures of crumb hardness have shown the closest correlation with the sensory per- ception of bread aging. (Axford and others 1968; Bashford and Hartung 1976; Gámbaro and others 2002). Consumers are the most appropriate tool for determining when a food product reaches the end of its shelf life. However, to repeatedly assemble consumer panels for the multiple measurements needed during shelf-life studies would be impractical and expensive. A sen- sory panel is more appropriate for repeated assessments, but it measures analytical attributes such as altered odor or hardness rather than directly assessing acceptability. How high does the altered odor measurement need to be for a trained panel’s assessment of product acceptability to decrease? The answer to this question can be ob- tained by correlating data obtained from a consumer panel with data obtained from a trained panel (Hough and others 2002). Correlations have been published for products such as sunflow- er kernels (Fritsch and others 1997), sunflower oil (Ramírez and oth- ers 2001), and milk powder (Hough and others 2002), but similar cor- relations have not been published for bakery products. Such corre- lations would be of enormous benefit for sensory quality control pro- grams. The use of survival analysis to study the shelf life of foods is quite a novel technique. Survival analysis, a branch of statistics, is extensively used in clinical studies, epidemiology, biology, sociol- ogy, and reliability studies (Kleinbaum 1996; Klein and Moesch- berger 1997; Meeker and Escobar 1998; Gómez and others 2001; Gómez 2002). Hough and others (2003) have introduced this meth- odology into the study of food shelf life. Their key concept is to fo- cus the shelf-life risk on the consumers’ rejection of the product rather than on product deterioration. The objectives of the study were (1) to correlate consumer re- sponse (global acceptability and acceptance/rejection data) to sen- sory and instrumental measurements, (2) to compare the accept- ability prediction models obtained in the 2 countries, and (3) to determine WPB shelf life by survival analysis methodology, on the basis of consumer acceptance/rejection of the product. Materials and Methods Samples The study was carried out independently in each country with WPB manufactured by local bakery plants. The printed commercial shelf life of products from both countries was 13 d. The ingredients used in the manufacture of the Spanish bread (provided by the manufacturer) were wheat flour, water, yeast, veg- etable oil, sugar, salt, powdered milk, fatty acid monoglycerides and diglycerides (E-471) and sodium stearoyl-2 lactylate (E-481), sodi- um propionate and potassium sorbate (E-281 and E-202), and vin- egar and guar gum (E-412). The ingredients used in the manufac- ture of the Uruguayan bread (provided by the manufacturer) were wheat flour, water, fructose, yeast, gluten, salt, honey, and calcium JFS S: Sensory and Nutritive Qualities of Food MS 20040282 Submitted 5/3/04, Revised 6/23/04, Accepted 8/27/04. Authors Gámbaro, Giménez, and Varela are with Sección Evaluación Sensorial, Cátedra de Ciencia and Tecnología de los Alimentos, Facultad de Química, Univ. de la República, Flores, Montevideo, Uruguay. Authors Fiszman and Salvador are with Inst. de Agroquímica y Tecnología de Alimentos (CSIC), Apartado de correos 73, Burjassot (Valencia), Spain. Direct inquiries to author Fiszman (E-mail: sfiszman@iata.csic.es).