Statistical analysis of the price and subjective quality ratings on Australian wines

Peter M. Visscher
DOI: https://doi.org/10.48550/arXiv.1402.3646
2014-02-15
Applications
Abstract:Consumers have a wide choice regarding the purchase of wine. Wines vary across varieties, regions, years of vintage, alcohol content, price and expert ratings. In this study we explored the relationship between those variables and whether a combination of explanatory variables can predict outcome variables using data on selected 2710 Australian wines rates by the latest book of James Halliday1. We analysed two dependent (outcome) variables, namely the rating and the price, and their (non-linear) relationship between explanatory variables and each other. We observe that the rating scale is not linear, as consumers may believe, and that a narrow range of rating scores is used for the ascertained wines. Across all wines, approximately 20% of variation in log(Price) can be explained by the explanatory factors state, variety, vintage and alcohol percentage. If the residuals from the model for rating is also taken as an explanatory variable for price, then a total of 53% of variation in log(Price) is explained by the model. For those wines ascertained to be in Halliday1, there is more variability between white wines than between red wines that is explained by explanatory variables such as State and variety. In comparison with hedonic pricing analyses in previous studies on French and Australian wines, a smaller proportion of variation in price and rating was explained in Australian wines that were in the 2014 edition of Halliday.
What problem does this paper attempt to address?