Most people would like to eat nothing but their favorite foods. Liking is a primary determinant of intake patterns and nutritional status, and therefore needs to be addressed in the study of patterns of human food intake. However, as adults, an increasing number of considerations other than liking influence food consumption patterns. At least three categories of factors influence adult intake: 1) concerns about procuring food such as costs and ease of obtaining and preparing the food, 2) concerns about the consequences of eating such as healthfulness, fat content, satiety value, and other anticipated consequences of ingestion, and 3) cultural rules about what constitutes food within the culture, rules of cuisine, and food taboos.
Concerns about procuring food, consequences of ingesting it, and knowledge of the cuisine rules of one's culture result from socialization, including the acquisition of information about food and eating, that continues throughout childhood. These concerns appear late in development relative to affective reactions to food, which the infant shows from birth. [1,2] In the early years of life, food likes and dislikes are the primary determinants of food intake.  Children are not yet influenced by many of the considerations that influence adult food acceptance patterns. However, likes and dislikes are not fixed, but are modified by early socialization and experience. Parental concerns regarding procuring food, consequences of ingesting it, and rules of cuisine have an indirect effect on children's food preferences and consumption patterns. [4,6] These parental concerns influence whether or not a food is made available to the child. This, in turn, influences frequency and quality of exposure, both important determinants of liking. Given the primacy of children's likes and dislikes, measures of preference can be especially useful as predictors of food consumption patterns. The child's affective response to food and how to measure it are the focus of this paper.
DEFINITION OF TERMS
Preference involves choice of one thing over others. In the strict behavioral sense, preference as choice implies nothing about the motivational process that leads to the choice. Defining preference in terms of behavioral choice has a history in the animal literature on learning and motivation.  A definition of preference as behavioral choice is, however, unnecessarily restrictive. It ignores verbal or gestural indicators of children's likes or dislikes, which often are given readily and spontaneously. For example, using a behavioral choice definition of preference, a child prefers liver over spinach, while other aspects of the child's behavior, including verbal responses and facial expression, reveal that both are disliked. For these reasons, a definition of preference that encompasses affect as a basis for choice is appropriate.
Throughout this review, preference refers both to 1) a continuum of hedonic response ranging from positive to negative, and 2) the relatively positive end of this continuum. Aversion refers to the negative end of the continuum. This approach moves beyond simple choice procedures, but makes measurement of preference less well defined. In measures of preference developed for children, the child is asked to make behavioral choices among foods based on affective reaction to the food. Because of the relatively direct relationship between preference and food intake in childhood,  information about children's food preferences should predict whether a food made available to a child will be consumed.
The preference assessment procedure was developed to obtain information from young children on their likes and dislikes for small numbers of foods. Most three-year-olds and nearly all older children can perform well in this task. Prior to actual assessment, performance of younger children can be …