=============================================================================== DATA DICTIONARY FOR FILE "PRICE_ZIP_DATA.CSV" By Rafael Becerril Arreola 02-95-2023 =============================================================================== The dataset price_zip_data.csv is released as companion to the paper "A Method to Assess and Explain Disparate Impact in Online Retailing", presented at the 2023 WebConf in Austin, TX. The paper describes the data collection process and the features of the main dataset. The variables included are as follows. zipcode: zip code where the product is to be delievered/shipped/picked up CBSA.TITLE: Core base statistical area of the zip code upc_recoded: numeric code for the UPC price: list price (before taxes) as displayed on website time: time the data point was collected PROPWHITE: proportion of white households in zip code PROPBLACK: proportion of black households in zip code PROPNATIVE: proportion of native households in zip code PROPASIAN: proportion of asian households in zip code PROPHISP: proportion of hispanic households in zip code AVGINC: average household income in zip code (in millions) To protect the identity of the retailer the data was collected from, UPC's corresponding to private labels have been removed. As a result, the number of observations in the data set is 528,111 but reduces to 527,523 observations after pairing as described in the paper. In addition, the UPC's have been recoded also to protect the identity of the retailer. The results with the reduced sample are qualitatively the same as those reported in the paper. With the reduce sample, the main regression table for the reduced dataset is as follows. ============================================================================================================ Dependent variable: ------------------------------------------------------------------------------------------------- Price difference None By product averaged By product,zip code averaged By product,zip code and time (w/i 60s) (1) (2) (3) (4) ------------------------------------------------------------------------------------------------------------ AVGINC 1.866 0.357*** 0.038** 0.030* (1.277) (0.073) (0.015) (0.016) PROPBLACK -0.495*** 0.099*** -0.011 -0.011 (0.171) (0.019) (0.008) (0.008) PROPHISP 0.210 0.125*** 0.021** 0.015** (0.319) (0.017) (0.008) (0.008) PROPASIAN 0.476 0.599*** 0.044*** 0.053*** (0.409) (0.034) (0.014) (0.015) PROPNATIVE 1.063 0.022 0.063*** 0.041* (0.663) (0.034) (0.024) (0.024) Constant 6.585*** 0.0003 0.003*** 0.001*** (0.160) (0.001) (0.001) (0.0005) -------------------------------------------------------------------------------------------------------------- Observations 527,523 527,523 527,523 527,523 R2 0.0002 0.010 0.0001 0.0001 Adjusted R2 0.0002 0.010 0.00005 0.0001 ============================================================================================================== Note: *p<0.1; **p<0.05; ***p<0.01