See the article "Time-series comparison of COVID-19 case fatality rates across 21 countries with adjustment for multiple covariates" on page 424.

I read the recently published article by Kim et al. [1]. On page 424 [1], the authors state, referring to my paper [2], that “other research using time-series cross-sectional data appears to have underestimated the impact of autocorrelation and heteroscedasticity”. However, this statement is incorrect and unfounded for 2 reasons. First, I used cross-sectional data rather than panel data, so there was no time component. The corollary is that residuals cannot be serially correlated. It makes no sense to consider autocorrelation in this case. Second, as shown in Section 5.1 of Perone [2], I safely considered heteroscedasticity in my paper: “Furthermore, since Breusch and Pagan (1979) and Shapiro and Wilk (1965) tests allowed to accept the null hypothesis of homoscedasticity and normality of residuals, models seemed well specified. However, due to the small sample, I preferred to adopt a conservative approach, by applying the HC2 correction proposed by MacKinnon and White (1985)” [3−5]. As a result, autocorrelation and heteroscedasticity issues have no bearing on the results of my paper.

Citations to this article as recorded by