Table 3.

Comparing Each Data Collection Group to Benford’s Law

Stock & inventoryFixed assetsProfitSales last week
(1)(2)(3)(4)
Panel A: Comparison of first digits across groups
Monthly = weekly (p)0.570.010.890.19
Weekly = phone (p)0.280.020.340.62
Monthly = weekly = phone (p)0.190.000.290.18
Panel B: Distance of first digit distribution from Benford’s Law
Monthly in-person0.070.120.070.06
Weekly in-person0.050.160.060.03
Weekly phone0.050.080.060.03
Sales last 4 weeksTotal costsMoney keptHousehold takings
(5)(6)(7)(8)
Panel A: Comparison of first digits across groups
Monthly = weekly (p)0.430.600.510.36
Weekly = phone (p)0.730.430.740.00
Monthly = weekly = phone (p)0.220.590.160.00
Panel B: Distance of first digit distribution from Benford’s Law
Monthly in-person0.040.050.130.17
Weekly in-person0.090.040.100.10
Weekly phone0.050.020.110.17
Stock & inventoryFixed assetsProfitSales last week
(1)(2)(3)(4)
Panel A: Comparison of first digits across groups
Monthly = weekly (p)0.570.010.890.19
Weekly = phone (p)0.280.020.340.62
Monthly = weekly = phone (p)0.190.000.290.18
Panel B: Distance of first digit distribution from Benford’s Law
Monthly in-person0.070.120.070.06
Weekly in-person0.050.160.060.03
Weekly phone0.050.080.060.03
Sales last 4 weeksTotal costsMoney keptHousehold takings
(5)(6)(7)(8)
Panel A: Comparison of first digits across groups
Monthly = weekly (p)0.430.600.510.36
Weekly = phone (p)0.730.430.740.00
Monthly = weekly = phone (p)0.220.590.160.00
Panel B: Distance of first digit distribution from Benford’s Law
Monthly in-person0.040.050.130.17
Weekly in-person0.090.040.100.10
Weekly phone0.050.020.110.17

Source: Authors’ analysis based on own data.

Note: This table compares distributions of first significant digits (FSDs). The first three rows report p-values from Wald tests that the distributions of FSDs are equal across data collection groups. These statistics are obtained by regressing indicators for each of the nine possible FSDs on group indicators in a system of equations, clustering standard errors by enterprise, and testing if the nine coefficients on each group indicator are jointly equal across groups. The final three rows report Euclidean distances (rescaled to be in the interval [0, 1]) between the observed FSD distribution for each data collection group and the distribution under Benford’s Law, following Cho and Gaines (2007).

Table 3.

Comparing Each Data Collection Group to Benford’s Law

Stock & inventoryFixed assetsProfitSales last week
(1)(2)(3)(4)
Panel A: Comparison of first digits across groups
Monthly = weekly (p)0.570.010.890.19
Weekly = phone (p)0.280.020.340.62
Monthly = weekly = phone (p)0.190.000.290.18
Panel B: Distance of first digit distribution from Benford’s Law
Monthly in-person0.070.120.070.06
Weekly in-person0.050.160.060.03
Weekly phone0.050.080.060.03
Sales last 4 weeksTotal costsMoney keptHousehold takings
(5)(6)(7)(8)
Panel A: Comparison of first digits across groups
Monthly = weekly (p)0.430.600.510.36
Weekly = phone (p)0.730.430.740.00
Monthly = weekly = phone (p)0.220.590.160.00
Panel B: Distance of first digit distribution from Benford’s Law
Monthly in-person0.040.050.130.17
Weekly in-person0.090.040.100.10
Weekly phone0.050.020.110.17
Stock & inventoryFixed assetsProfitSales last week
(1)(2)(3)(4)
Panel A: Comparison of first digits across groups
Monthly = weekly (p)0.570.010.890.19
Weekly = phone (p)0.280.020.340.62
Monthly = weekly = phone (p)0.190.000.290.18
Panel B: Distance of first digit distribution from Benford’s Law
Monthly in-person0.070.120.070.06
Weekly in-person0.050.160.060.03
Weekly phone0.050.080.060.03
Sales last 4 weeksTotal costsMoney keptHousehold takings
(5)(6)(7)(8)
Panel A: Comparison of first digits across groups
Monthly = weekly (p)0.430.600.510.36
Weekly = phone (p)0.730.430.740.00
Monthly = weekly = phone (p)0.220.590.160.00
Panel B: Distance of first digit distribution from Benford’s Law
Monthly in-person0.040.050.130.17
Weekly in-person0.090.040.100.10
Weekly phone0.050.020.110.17

Source: Authors’ analysis based on own data.

Note: This table compares distributions of first significant digits (FSDs). The first three rows report p-values from Wald tests that the distributions of FSDs are equal across data collection groups. These statistics are obtained by regressing indicators for each of the nine possible FSDs on group indicators in a system of equations, clustering standard errors by enterprise, and testing if the nine coefficients on each group indicator are jointly equal across groups. The final three rows report Euclidean distances (rescaled to be in the interval [0, 1]) between the observed FSD distribution for each data collection group and the distribution under Benford’s Law, following Cho and Gaines (2007).

Close
This Feature Is Available To Subscribers Only

Sign In or Create an Account

Close

This PDF is available to Subscribers Only

View Article Abstract & Purchase Options

For full access to this pdf, sign in to an existing account, or purchase an annual subscription.

Close