A Latent Look at Empirical Generalizations

All empirical data and the resulting parameters are subject to error. In this paper we explicitly use the standard, but hardly profound, "model" $$\hbox{Observed Value} = \hbox{True Score} + \hbox{Error}$$ as a lens for better viewing empirical studies in search of empirical generalization...

Full description

Saved in:
Bibliographic Details
Published inMarketing science (Providence, R.I.) Vol. 14; no. 3_supplement; pp. G61 - G70
Main Authors Morrison, Donald G., Silva-Risso, Jorge
Format Journal Article
LanguageEnglish
Published Linthicum INFORMS 01.08.1995
Institute for Operations Research and the Management Sciences
SeriesMarketing Science
Subjects
appropriate comparisons
Brand preferences
Brands
Consumer behavior
Consumers
Estimators
Generalizations
Households
Marketing
Marketing mixes
Mathematical models
Maximum likelihood estimation
Price elasticity
Probability distribution
Purchasing
Science
shrinkage estimators
Standard error
Statistical analysis
Statistical data
Statistical variance
Stochastic Choice Phenomena
summary statistics
unobservable constructs
Unobservables
Variance
United States > US
Online AccessGet full text
ISSN0732-2399
1526-548X
DOI10.1287/mksc.14.3.G61

Cover

More Information
Summary:All empirical data and the resulting parameters are subject to error. In this paper we explicitly use the standard, but hardly profound, "model" $$\hbox{Observed Value} = \hbox{True Score} + \hbox{Error}$$ as a lens for better viewing empirical studies in search of empirical generalizations in marketing. This lens is especially valuable when the unit of analysis is the individual consumer. However, even when a macro study contrasts price elasticities across cities, this O = T + E framework can be very helpful.
Bibliography:SourceType-Scholarly Journals-1
ObjectType-Statistics/Data Report-1
content type line 14
ObjectType-Article-1
ObjectType-Feature-2
ISSN:0732-2399
1526-548X
DOI:10.1287/mksc.14.3.G61