search query: @author Oumlil, B. / total: 1
reference: 1 / 1
« previous | next »
| Author: | Oumlil, B. Balloun, J. L. |
| Title: | An alternative approach to developing multiple regression models for marketing survey research |
| Journal: | Journal of the Market Research Society
1990 : OCT, VOL. 32:4, p. 521-535 |
| Index terms: | MULTIPLE REGRESSION ANALYSIS MULTIVARIATE ANALYSIS MARKETING RESEARCH MARKET SURVEYS |
| Language: | eng |
| Abstract: | Multiple regression analysis is the most frequently used multivariant technique in marketing research. However, there are some challenges, that arise because the classical assumptions are often violated. The stepwise technique can yield biased results /large type I errors, model misspecification/, and is not to be considered a method for reaching conclusions. The paper demonstrates the applicability of an alternative method for developing multiple regression models for marketing panel data. It is compared with the classical stepwise, forward multiple regression /SFMR/ technique, provides more conservative estimates /ie. fewer "significant" results/ and less biased estimates than SFMR analysis. |
« previous | next »
SCIMA