C12 - Hypothesis Testing: GeneralReturn

The first large-scale (909 products, 79,679 individual price listings) empirical study of price dispersion based on data from the Czech Republic's online shops is presented in the paper. First, simple descriptive indicators were calculated for each product. Second, several versions of linear regression models were constructed for each of the 13 product categories and evaluated against the hypotheses. The price dispersion was regressed against its mean market price and the number of shops. For the majority of categories R² is high or very high (the mean market price variable accounts for 86.6% to 96.7% in the price dispersion) or at least medium (from 56% to 69%). Also, price dispersion (measured by range and standard deviation) remains quite high for the majority of the analyzed product categories. The important role of pricebots for online shop owners and marketing managers and some other findings are also discussed.