Title: Robustness and sensitivity of conjoint analysis versus multiple linear regression analysis

Authors: Fahri Karakaya; Abhrawashyu Awasthi

Addresses: University of Massachusetts Dartmouth, Charlton College of Business, Department of Marketing, Dartmouth, MA 02747-2300, USA ' Dr. Fresh LLC, 6645 Caballero Blvd, Buena Park, CA 90620, USA

Abstract: This study compares the robustness of conjoint analysis versus multiple linear regression when using orthogonal data. The explained variance (R²) by four independent variables was utilised to test the robustness of the regression analysis while Pearson's R and Kendall's tau were used for testing conjoint method. The results indicate that the two methods produce somewhat different results and conjoint analysis is more robust compared to regression.

Keywords: conjoint analysis; sensitivity analysis; multiple linear regression analysis; robust statistics; orthogonal data; robustness.

DOI: 10.1504/IJDATS.2014.062461

International Journal of Data Analysis Techniques and Strategies, 2014 Vol.6 No.2, pp.121 - 136

Available online: 08 Jun 2014 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article