统计学家纳西姆·塔勒布对一项美国政府委托的酒精摄入健康研究提出质疑[1]。这项发表于《酒精与药物研究期刊》2026年第87卷第4期的研究声称，每周饮酒7杯时相关死亡风险超过千分之一，14杯时达到约4%，并建议每日饮酒量不超过1杯[1]。然而塔勒布指出，该研究的统计基础存在多个问题，不足以支撑推翻数十年饮酒建议的结论[1]。
塔勒布的批评集中在三个方面[1]。首先，研究使用了不透明的非对称置信区间计算方法，以每周7杯饮酒且年龄超过70岁的人群为例，其数值-5.47明显不在报告的置信区间[-28.59, 14.57]的中点范围内[1]。其次，总体风险区间（31.86）紧于独立求和（35.81）和二次方和（22.42），但研究未说明如何处理数据间的相关性[1]。第三，年长人群数据对结果的影响过大，超过70岁人群的数据在男性21杯饮酒风险中贡献了27.30/68.92，表明研究过度依赖高方差的老年人群组数据[1]。
Nassim Taleb has challenged the methodology of a U.S. government-commissioned study on alcohol consumption and health, arguing that its statistical foundation is insufficient to support its major conclusions [1]. The study, published in the Journal of Studies on Alcohol and Drugs in 2026 (Volume 87, Issue 4), claims that moderate to low alcohol consumption provides no protective health effects and recommends limiting daily intake to no more than one drink [1].
According to the research, the mortality risk associated with alcohol consumption exceeds 1 per 1,000 at seven drinks per week and rises to approximately 4 percent at 14 drinks per week [1]. However, Taleb identifies several methodological problems undermining these findings. He points out irregularities in the confidence interval calculations, particularly noting that a value of -5.47 for men over 70 consuming seven drinks weekly falls outside its reported interval of [-28.59, 14.57] [1]. Additionally, Taleb contends the study employs non-transparent asymmetric confidence intervals and improper risk aggregation methods [1].
A significant concern Taleb raises involves the study's reliance on data from elderly populations. For men consuming 21 drinks per week, data from those over 70 years old contributes 27.30 out of a total 68.92 risk points, yet the study does not adequately explain how it accounts for the correlation between risk factors across different age groups [1]. The overall risk interval reported (31.86) is narrower than both the independent sum (35.81) and the quadratic sum (22.42), raising questions about the treatment of correlations [1]. Taleb maintains that these statistical limitations undermine the study's ability to overturn decades of established guidance on alcohol consumption [1].