33:630:385 Lecture Notes - Lecture 10: Multinomial Logistic Regression, Dependent And Independent Variables, Exponential Function
18 Apr 2016
School
Department
Course
Professor
Document Summary
Limitaions: requires respondent to think about each atribute individually, determining importance of each atribute oten diicult task, paricularly as # of atributes increases. An alternate approach than asking than asking individual atributes. Based on premise that product is a bundle of atributes. Respondent rates a set of compeing products in product category. Researcher uses regression to determine how much each atribute impacts overall raing. Step 1: idenify key product atributes and levels of atributes for the product class: batery life, exterior, price: Step 2: make hypotheical products using combinaions of atributes and atribute levels. Step 3: respondents provide a raing for each hypotheical product. Step 4: researcher analyzes raings data: create dummy variables for atribute levels, using dummy variables, esimate: Raing = a + (b1*batery_10) + (b2*batery_12) + (b3*silver) + (b4*price_1299) Use part-worths to predict raing for any hypotheical product.
