Proceedings

Consumer attitudes to health and pleasure - survey data studied by PLSR.

Year	2006
Abstract	The paper demonstrates how Partial Least Squares regression (PLSR) can be used to handle consumer survey data in a rapid, statistical reliable and graphical efficient way. A representative group of Danish consumers (n=1655) responded to a questionnaire including 12 questions concerning attitudes to health and pleasure. Furthermore, various demographic variables were collected from the consumers. Two linear models Y=XB+F were established by PLSR, using the questionnaire data as 12 Y-variables in both models. Variables segmenting the consumers in two different ways, based on the demographic and attitudes data, constituted the Xvariables. The models were validated by jack-knifing and results presented as plots of the combined significance level and sign of the regression coefficients from the PLSR models at optimal rank (5 PLS components). Relationships between consumer questionnaire responses and their background (demographic variables or the value-based segments) could then be studied.
Reference	Martens, M., Frøst, M.B., Martens, H. 2006. Consumer attitudes to health and pleasure - survey data studied by PLSR. Proc. PLS 2005, 2005, PLS and related methods,,,,,,,,,,,,, Barcelona, Spain, pp 431-437.
Publisher	PLS and related methods,,,,,,,,,,,,,