Similarities and differences in brand purchase behavior across categories [electronic resource] / Andrew Scott Ainslie.

Ainslie, Andrew Scott.
Bib ID
vtls001055185
出版項
Ann Arbor, Mich. : ProQuest Information and learning
稽核項
100 p.
電子版
預約人數:0
全部評等: 0
沒有紀錄。
 
 
 
03243nam a2200313 a 4500
001
 
 
vtls001055185
003
 
 
VRT
005
 
 
20071225064200.0
006
 
 
m        d 
008
 
 
071225s1998    miu||||||m   |000 0|eng d
020
$a 9780599093072
035
$a (UMI)AAI9910950
039
9
$y 200712250642 $z VLOAD
040
$a UMI $b eng $c UMI
100
1
$a Ainslie, Andrew Scott.
245
1
0
$a Similarities and differences in brand purchase behavior across categories $h [electronic resource] / $c Andrew Scott Ainslie.
260
$a Ann Arbor, Mich. : $b ProQuest Information and learning
300
$a 100 p.
500
$a Source: Dissertation Abstracts International, Volume: 59-11, Section: A, page: 4220.
500
$a Adviser: Peter E. Rossi.
502
$a Thesis (Ph.D.)--The University of Chicago, 1998.
520
$a Models of brand choice have generally focused on one category, ignoring the fact that purchase behavior data for panel households is often available across multiple categories. This dissertation develops new models that allow one to analyze both similarities and differences in heterogeneity exhibited by consumers across different categories. In Chapters 2 and 3, a multi-category Bayesian hierarchical Probit model of consumer choice conditional upon a purchase being made is explored. It separates out the within-category and between-category effects, by placing restrictions on the covariance matrix of the parameters on the marketing mix variables. The model allows analysis of the differences in behavior for households based on their demographics and shopping behavior.
520
$a Two important findings emerge. For a given household, a far higher proportion of behavior is shown to be common across categories than in other research to date. Also, two different components of shopper "expertise" are found to operate in opposing directions. Shoppers who purchase more frequently overall are more price sensitive; yet heavy users of a particular category are less price sensitive in that category.
520
$a This first model only considers behavior conditional upon the household's decision to purchase in the category (the "conditional" model). In the second model, developed in chapter 4, the decision process is expanded to consider the choice to purchase in the category, and the quantity purchased (the "unconditional" model). This permits the analysis of correlations in coefficients on variables that affect the household's threshold price to purchase across categories. It is a departure from previous research in that it allows a far more flexible distribution of the error terms on the latent utilities. We show that the GEV assumption of a single variance scale parameter across all error terms is overly restrictive. In the variance components model where information is combined across categories, we find that including the additional information made available from considering non-purchase occasions is useful in determining what drives the household's behavior.
653
$a Business Administration, Marketing.
710
2
0
$a The University of Chicago.
773
0
$t ABI/INFORM Global (ProQuest) $g 59-11A.
856
4
1
$u http://info.lib.tku.edu.tw/ebook/redirect.asp?bibid=1055185
999
$a VIRTUA00
沒有評論
摘要
Models of brand choice have generally focused on one category, ignoring the fact that purchase behavior data for panel households is often available across multiple categories. This dissertation develops new models that allow one to analyze both similarities and differences in heterogeneity exhibited by consumers across different categories. In Chapters 2 and 3, a multi-category Bayesian hierarchical Probit model of consumer choice conditional upon a purchase being made is explored. It separates out the within-category and between-category effects, by placing restrictions on the covariance matrix of the parameters on the marketing mix variables. The model allows analysis of the differences in behavior for households based on their demographics and shopping behavior.
Two important findings emerge. For a given household, a far higher proportion of behavior is shown to be common across categories than in other research to date. Also, two different components of shopper "expertise" are found to operate in opposing directions. Shoppers who purchase more frequently overall are more price sensitive; yet heavy users of a particular category are less price sensitive in that category.
This first model only considers behavior conditional upon the household's decision to purchase in the category (the "conditional" model). In the second model, developed in chapter 4, the decision process is expanded to consider the choice to purchase in the category, and the quantity purchased (the "unconditional" model). This permits the analysis of correlations in coefficients on variables that affect the household's threshold price to purchase across categories. It is a departure from previous research in that it allows a far more flexible distribution of the error terms on the latent utilities. We show that the GEV assumption of a single variance scale parameter across all error terms is overly restrictive. In the variance components model where information is combined across categories, we find that including the additional information made available from considering non-purchase occasions is useful in determining what drives the household's behavior.
附註
Source: Dissertation Abstracts International, Volume: 59-11, Section: A, page: 4220.
Adviser: Peter E. Rossi.
Thesis (Ph.D.)--The University of Chicago, 1998.
合著者
ISBN/ISSN
9780599093072