关于方差分析后组间比较的问题

sanming_1978

文章修回，关于方差分析的问题，请大牛指点。
一是，reviewer提出的问题是什么？我不是很明白；二是、我的统计分析怎么完善？ 谢谢！

reviewer 1 的问题：

If the authors wish to perform all the statistical analyses on all peak measures separately, they should simply apply corrections for multiple comparisons.
然而，这些只是建议。如果作者希望对所有的峰值措施分别进行所有的统计分析，他们应该简单地应用多个比较的校正。

The statistics section does not mention tests or corrections for sphericity, which should be included. Also, the results section does not report degrees of freedom for the F values in the ANOVA – these should also be reported.
统计部分没有提到球形检验或修正，应该包括球形检验。此外，结果部分没有F值的自由度，在方差分析-这些也应该报告。

The ANOVA results are currently reported only in a table, but not given in the text of the results section where the effects are discussed. This is awkward for the reader, having to switch between text and table to assess the effects and statements. The F and p values (with degrees of freedom included) should be reported in the results text, at the point where the statements about the effects are made.
方差分析结果目前只报告在一个表中，但没有给出的结果部分的文本中的影响进行了讨论。这是对读者来讲很尴尬，必须切换文本和表之间的效果和语句的评估。应该在文本结果中说明F和P值（包括自由度），并说明影响。

reviewer 2 的问题：

The authors analyzed peak latencies, peak amplitudes and V-A slope by first conducting 13 separate 2 X 2 (Age X Sex) ANOVAs (one for each measure, as reported in Table 1). Next, they conducted all possible pairwise comparisons using simple independent measures t-tests (reported in Table 2), thereby ignoring the outcomes of the overall ANOVAs.
作者分析了峰潜伏期、波幅和V-A斜率进行13个独立的2×2（年龄×性别）方差分析（每一个测试均报道在了表1中）。接下来，你们做了所有可能的两两比较，采用简单的独立样本t检验的措施（见表2），从而忽略总体的方差分析结果。【这个可能是主要点，不知道如何评价整体性】

Conducting overall 2 X 2 ANOVAs with pair-wise comparisons FOR SIGNIFICANT INTERACTION EFFECTS would have been a reasonable approach to statistical analysis; however, this is not actually what was done. 成对比较显著的交互效应进行整体2×2方差分析将是统计分析的一个合理的方法；然而，这是不是真的做到了。

Instead of conducting follow up pairwise comparisons (young vs. adult for males and females, and females vs. males for children and adults) for only those measures for which the Sex X Age interaction was significant, the authors ignored the outcome of the ANOVAs and simply performed all possible pairwise comparisons using simple t-tests.
作者没有跟进两两比较（年轻与成年男性和女性，女性与儿童和成人男性），分析只有那些措施的性X年龄的交互作用显著，因此作者忽略了方差分析的结果，简单地对所有可能的成对比较采用了简单使用t检验。

For measures without significant interaction effects in the overall ANOVA, significant main effects of Age (children vs. adults) and Sex (female vs. male), when present, are unambiguous and simple to interpret, and other follow-up t-tests (e.g., boys vs. adult women) are not warranted. Thus, for the amplitude analysis, only the amplitude of Wave A would need a follow-up, since this was the only wave for which the Sex X Age interaction was significant.
因为没有总体方差显著交互效应的情况下，当出现时显著的主效应年龄（儿童和成人）和性别（男、女）影响时，是明确的解释。与其他后续检验（例如，男孩和成年女性）是需要。因此，对于波幅分析，只有波A的振幅需要后续检验，因为这是唯一一个性别X年龄的相互作用是显着的波。

The authors should re-evaluate their statistical approach. The value of using a two-way ANOVA is that it shows interaction effects; and when interactions are significant, then the main effects cannot be interpreted simply (without qualification of the second factor) as the authors have attempted to do.
作者应该重新评估他们的统计方法。使用双向方差分析显示了相互作用影响时，当相互作用是显着的，那么主要的影响不能被简单地解释，作者试图做。

sanming_1978

为什么看不到图片

abc886y365hxg

别人叫你进行校正，还有就是检验方法可能有问题.现在回复你可能晚了，因为前段时间有事没上线，如果还有需要，可以联系我QQ710756196