In Arslan et al. (2018), we reported ovulatory increases in extra-pair sexual desire, in-pair sexual desire, and self-perceived desirability, as well as several moderator analyses related to the good genes ovulatory shift hypothesis, which predicts attenuated ovulatory increases in extra-pair desire for women with attractive partners. Gangestad and Dinh (2021) identified errors in how we aggregated two of the four main moderator variables. We are grateful that their scrutiny uncovered these errors. After corrections, our moderation results are more mixed than we previously reported and depend on the moderator specification. However, we disagree that the evidence for moderation is robust and compelling, as Gangestad and Dinh (2021) claim. Our data are consistent with some previously reported effect sizes, but also with negligible moderator effects. We also show that what Gangestad and Dinh (2021) call an "a priori[…]more comprehensive and valid composite" is poorly justifiable on a priori grounds, and follow-up analyses they report are not robust to a composite specification that we consider at least as reasonable. Psychologists have to become acquainted with techniques such as cross-validation or training and test sets to manage the risks of data-dependent analyses. In doing so, we might learn that we need new data more often than we intuit and should remain uncertain far more often. (PsycInfo Database Record (c) 2021 APA, all rights reserved).

中文翻译：

---

Arslan 等人中良好基因排卵变化的证据。(2018) 是混合和不确定的。

在阿尔斯兰等人。（2018 年），我们报告了成对额外性欲、成对性欲和自我感知欲望的排卵增加，以及与好基因排卵转移假设相关的几种调节分析，该假设预测了额外排卵期的排卵增加减弱。将女性与有吸引力的伴侣配对。Gangestad 和 Dinh（2021 年）发现了我们汇总四个主要调节变量中的两个的错误。我们很感激他们的审查发现了这些错误。更正后，我们的审核结果比我们之前报告的要复杂得多，并且取决于审核员的规范。然而，正如 Gangestad 和 Dinh（2021）所声称的那样，我们不同意适度的证据是强有力的和令人信服的。我们的数据与之前报道的一些效应大小一致，但调节剂效应可以忽略不计。我们还表明，Gangestad 和 Dinh（2021 年）所谓的“先验[……]更全面和有效的复合”在先验基础上是不合理的，并且他们报告的后续分析对于我们认为的复合规范并不稳健至少是合理的。心理学家必须熟悉交叉验证或培训和测试集等技术，以管理依赖数据的分析的风险。这样做时，我们可能会了解到，我们需要新数据的频率比我们的直觉要高得多，并且应该更频繁地保持不确定性。（PsycInfo 数据库记录 (c) 2021 APA，保留所有权利）。他们报告的后续分析对于我们认为至少是合理的复合规范并不可靠。心理学家必须熟悉交叉验证或培训和测试集等技术，以管理依赖数据的分析的风险。这样做时，我们可能会了解到，我们需要新数据的频率比我们的直觉要高得多，并且应该更频繁地保持不确定性。（PsycInfo 数据库记录 (c) 2021 APA，保留所有权利）。他们报告的后续分析对于我们认为至少是合理的复合规范并不可靠。心理学家必须熟悉交叉验证或培训和测试集等技术，以管理依赖数据的分析的风险。这样做时，我们可能会了解到，我们需要新数据的频率比我们的直觉要高得多，并且应该更频繁地保持不确定性。（PsycInfo 数据库记录 (c) 2021 APA，保留所有权利）。