Jay Beaman, Jerry Vaske, Michel Grenier
This article identifies the need to consider the consequences of response heaping in making estimates from data in which frequency functions of stated responses show that some numeric responses (such as numbers ending in 0 or 5) are preferred. A framework is introduced for recognizing heaping on certain numbers (number preference) as unique and separable from recall bias or other biases that result in stated responses being precise, but systematically different from accurate or factually correct responses. A prototype model for heaping and determining bias effects is presented and illustrated using travel and tourism data. Concerns are raised about whether results derived are just precise or accurately estimate actual behavior.
© 2001-2024 Fundación Dialnet · Todos los derechos reservados