Where does attribute collide?

In metaphysics an attribute is a property predicated of a substance, a feature by which a thing is characterised and known. Spinoza's 1677 Ethics defined an attribute as what the intellect perceives of a substance as constituting its essence, and gave thought and extension as the two attributes through which the one substance is grasped; the idea reaches back to Aristotle's account of what is predicated of a subject. Here the referent is an ontological feature of being itself, argued for and defined, not a column in a stored table, a countable quality of a manufactured item, or a product feature a shopper weighs.[1]

Unit

One property predicated of a substance, such as thought or extension

Recovery means

Not applicable

Measured by

Argued for by definition and inference, and identified through what the intellect perceives of a substance rather than measured empirically

In marketing and consumer choice a product attribute is a characteristic of an option, such as price, size, or a specific feature, on which buyers form beliefs and which they weigh when choosing. Fishbein's 1963 study of beliefs about an object modelled an attitude as the sum of a person's beliefs about the object's attributes weighted by how each attribute is evaluated, and Green and Rao's 1971 conjoint measurement made attribute trade-offs estimable from choice data. Here the referent is a perceived feature of a purchasable option that enters a decision, not a property of substance, a countable quality of production, or a named column of stored data.[2][3]

Unit

One perceived feature of a product option, on a rated or traded-off scale

Recovery means

Not applicable

Measured by

Elicited by rating scales, multiattribute attitude models, and conjoint analysis of choice trade-offs

In the relational model of databases an attribute is a named property of a relation, a column that draws its values from a defined domain and, together with the other attributes, gives each stored tuple its structure. Codd's 1970 A relational model of data for large shared data banks defined a relation as a set of tuples over named attributes, and Chen's 1976 entity-relationship model carried the term over to the properties of entities and relationships in data design. Here the referent is a typed, named slot in a schema populated with values, not a property of being, a categorical quality in quality control, or a feature a consumer evaluates.[4][5]

Unit

One named column of a relation, valued from a defined domain

Recovery means

Not applicable

Measured by

Declared in a schema with a name and a domain, and populated with a value in each tuple