|
Activity Number:
|
443
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 6, 2008 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Statistics and Marketing
|
| Abstract - #300937 |
|
Title:
|
Determining the Impact of Attitudinal Drivers on Customer Satisfaction in the Presence of Missing Data
|
|
Author(s):
|
Kurt A. Pflughoeft and Jorge A. Alejandro*+
|
|
Companies:
|
Market Probe and Market Probe
|
|
Address:
|
2655 N Mayfair Rd, Milwaukee, WI, 53226,
|
|
Keywords:
|
Missing Data ; Multiple Imputation ; Attitudinal Drivers ; Relative Importance ; Random Forest ; MARS
|
|
Abstract:
|
With the advent of the Multi-Attribute Attitudinal Model in the early 1960's, researchers have been concerned not only with measuring attitudes related to customers satisfaction but also with determining the importance of each attribute. Estimating the importance of attitudes on overall satisfaction is a vexing problem and missing values can only further exacerbate it. Multiple imputation is used in conjunction with several importance techniques which measure either predictive accuracy or overall information content. A simulation study is presented whereby missingness is artificially induced in a Missing At Random (MAR) fashion in order to determine how the effect size of importance measures is affected by various missing value techniques. Importance estimates are derived using MARS, Random Forests and Theil's relative importance.
|
