The Impact Factor (IF), despite its widespread use, suffers from well-known biases that remain incompletely addressed in practice -- most notably its sensitivity to journal size and its lack of field normalization. Because of size sensitivity, a randomly formed journal of $n$ papers can attain a range of IF values that decreases sharply with size, as $\sim 1/\sqrt{n}$. The Central Limit Theorem, which underlies this effect, also allows us to correct for it by standardizing citation averages for scale and field in a manner analogous to calculating the $z$-score in statistics. We thus introduce the $Φ$ (Phi) index, defined as $Φ= (f - μ)\sqrt{n}/σ$, where $f$ is a journal's average citation count (akin to the IF), $n$ its publication count, and $μ, σ$ the mean and standard deviation of citations in its field. Applying the $Φ$ index to 12,173 journals in Clarivate's Journal Citation Reports, we obtain rankings that correct for size bias and elevate journals from underrepresented fields such as mathematics, law, and history. We validate the $Φ$ index via a Monte Carlo random sample test, which we propose as a standard diagnostic for any citation indicator. The methodology extends readily to departments, universities, and countries.

翻译：影响因子（IF）虽被广泛使用，却存在尚未在实践中完全解决的显著偏差——最突出的是其对期刊规模的敏感性以及缺乏领域归一化。由于规模敏感性，一个由 $n$ 篇论文随机组成的期刊，其可能的影响因子值范围随规模锐减，变化关系为 $\sim 1/\sqrt{n}$。支撑这一效应的中心极限定理，也使我们能够通过类似统计学中计算 $z$ 分数的方式来标准化引文平均值的尺度和领域，从而纠正该偏差。由此我们引入 $Φ$（Phi）指数，定义为 $Φ= (f - μ)\sqrt{n}/σ$，其中 $f$ 为期刊的平均引文数（类似于IF），$n$ 为其论文总数，$μ$ 和 $σ$ 分别为其领域内引文的均值与标准差。我们将 $Φ$ 指数应用于科睿唯安《期刊引证报告》中的 12,173 种期刊，所得排名纠正了规模偏差，并提升了数学、法律、历史等代表性不足领域期刊的地位。我们通过蒙特卡罗随机样本检验验证了 $Φ$ 指数，并将其推荐为任何引文指标的标准诊断方法。该指标可轻松推广至院系、大学和国家层面。