When exploring decision making in food marketing, particular events are often unequally represented in the data. Concerns over bias can confuse public policy or strategy interpretation if samples are unbalanced. This occurs when researching organic food marketing. A small share of the market (∼4%) and of product innovations (∼7%) are organic. Caution is advised when predicting trends based on this disproportionately small sample of organic products. This study demonstrates that there is little statistical difference, and even a net gain in predictive power, when using a balanced sample to test factors that influence a firm’s decision to market organic food. Marginal probabilities for the smaller, balanced sample are biased by a constant factor. Thus ex post adjustment is easy and recommended to enhance predictive power.
Shanahan, Christopher J. and Neal H. Hooker. 2016. A Balancing Act: Disproportionate Sampling of Organic Foods. Journal of Food Products Marketing. 22(3): 320‐331