search query: @indexterm CONSUMER RESEARCH / total: 600
reference: 112 / 600
| Author: | Teichert, T. |
| Title: | Schätzgenauigkeit von Conjoint-Analysen |
| Journal: | Zeitschrift für Betriebswirtschaft
1998 : VOL. 68:11, p. 1245-1266 |
| Index terms: | CONJOINT ANALYSIS CONSUMER RESEARCH ESTIMATION PREFERENCES |
| Language: | ger |
| Abstract: | Conjoint analysis is widely applied in the area of consumer research to estimate metric preference functions from ranking input data. This paper shows the potential pitfalls which stem from the necessary scale transformation. It is shown that different estimation methods lead to divergent solutions. The part-worth estimates possess estimation intervals of remarkable size. The estimation intervals are indeterminate both in their size and location and depend on experimental design, utility function and error term. The commonly used estimation methods OLS and LINMAP provide only suboptimal estimates and can lead to an exaggeration of individual differences in utility functions. Potential for improvement is shown. |
SCIMA