The authors challenge Barabási's claim that human email communication patterns follow a power-law distribution with exponent α ≈ 1. They argue that the reported power-law is an artifact of the data analysis and that the actual distribution is better described by a log-normal distribution. Barabási analyzed email data from 3188 individuals over 83 days, but the authors found significant flaws in his analysis, including an unrealistic assumption that very short time intervals between emails are possible. They also show that the data are better described by a log-normal distribution, which is common in real-world processes. A Bayesian model selection analysis supports the log-normal model, as it has a higher posterior probability. The authors also critique Barabási's priority-queuing model, which predicts a power-law distribution for email response times, but this is not supported by the data. The model predicts unrealistic behavior, such as a peak at τ = 1 second and a uniform distribution of task priorities. The authors conclude that the reported results in Barabási's study do not hold up under further inspection. The study highlights the importance of careful data analysis and the need to consider alternative models when interpreting empirical data.The authors challenge Barabási's claim that human email communication patterns follow a power-law distribution with exponent α ≈ 1. They argue that the reported power-law is an artifact of the data analysis and that the actual distribution is better described by a log-normal distribution. Barabási analyzed email data from 3188 individuals over 83 days, but the authors found significant flaws in his analysis, including an unrealistic assumption that very short time intervals between emails are possible. They also show that the data are better described by a log-normal distribution, which is common in real-world processes. A Bayesian model selection analysis supports the log-normal model, as it has a higher posterior probability. The authors also critique Barabási's priority-queuing model, which predicts a power-law distribution for email response times, but this is not supported by the data. The model predicts unrealistic behavior, such as a peak at τ = 1 second and a uniform distribution of task priorities. The authors conclude that the reported results in Barabási's study do not hold up under further inspection. The study highlights the importance of careful data analysis and the need to consider alternative models when interpreting empirical data.