Table 3 First digit tests. The table presents the results of estimating Eq. 6 using OLS.

From: Using the Newcomb–Benford law to study the association between a country’s COVID-19 reporting accuracy and its development

Variable

Confirmed cases

Death cases

(1)

(2)

Panel A. 185 countries

Developmental index

\(- 0.56^{***}\)

( 0.00)

\(-1.83^{***}\)

( 0.00)

ln(Population)

\(-1.38^{***}\)

( 0.00)

\(-4.25^{***}\)

( 0.00)

No. of days

\(-0.78^{***}\)

( 0.00)

\(-2.30^{***}\)

( 0.00)

Sample size

154

139

\({Adj. R}^{2}\)

13.92%

34.07%

Panel B. 50 countries with regional data

Developmental index

\(-0.39^{*}\)

( 0.07)

\(-2.38^{**}\)

( 0.02)

ln(Population)

\(-0.25\)

( 0.32)

\(-2.77^{*}\)

( 0.10)

No. of days

\(-0.02^{***}\)

( 0.00)

\(-0.13^{***}\)

( 0.00)

Sample size

49

29

\({Adj. R}^{2}\)

13.77%

24.68%

  1. The dependent variable is the D goodness-of-fit measure for first digits. The unit of observation is a country. Panel A shows the results for the whole dataset with 185 countries for the cumulative number of confirmed cases and death cases; panel B shows the results for 50 countries with regional data. To avoid small coefficients, we divide the development index and ln(Population) by 100 and No. of Days by 1000 for all models. Sample sizes vary due to missing values. P-values for a one-tailed t-test are in parentheses. ***, **, and * denote significance at the 1%, 5% and 10% levels, respectively.