我的稿件因研究“功效不足”而被退回。这是什么意思？我已经采用了最佳的研究方法。 在统计学里，“功效”指的是您的研究是否能鉴定具有重要权益的效应。基本上，进行研究设计时，必须考虑到以下四个必要的因素： 1. 样本数：单位（例如，病人）的数目，通常以“N”做代表。 2. 研究效应的大小：一般而言，若想达到的效应越大，所需的样本量相对较小。 3. α水平：统计意义的阈值（可定为 .001 、 .05 或 .1 ）。当数据的p值等于或超越此临界值时，代表您的研究成果不具任何统计意义。 4. 功效：这是个数值，代表您能得到某个效应的可能性。 该怎么确定你研究的功效？以上列出的四个因素是相互关联的，若你有其中三个因素的数值，就能计算出第四个因素的数值。通常α水平是固定的（你得在 .001 、 .05 或 .1 之间选其一），在查阅相关文献后，你对研究效应的大小会有个概念。若想让研究取得有力的功效，就得关注样本数的多寡。 多数权威期刊如《自然》（ Nature ）都会要求对研究制定的样本量做出解释，以证明成果带有足够的功效。《自然》也提供了具体的 指导方针 ，建议在研究的样本量小的情况下应该进行那些测试。其他期刊如《 英国外科学杂志 》（ British Journal of Surgery ）指定稿件必需包含明确的功效计算法。有些期刊如《分子遗传学和新陈代谢杂志》（ Molecular Genetics and Metabolism ）更直接的表明：“ 递交的稿件若没有附加功效的计算，将一概被拒绝，并在未经审核的情况下退还给作者。 ”除了跟医学和生命科学有关的期刊外，其他类型的期刊也同样对统计功效有着同样严格的要求。比如，美国心理学会发表的《 心理学研究报告准则 》（ Reporting Standards for Research in Psychology ）就强力的推荐作者于稿件内的研究方法部分阐明对功效的分析。 在申请研究基金时，若能把功效的计算包括在申请书里，能帮助评审评估研究的可行性。 相信你已经注意到以上的说明并没不包括研究所采用的方法。这是因为功效与研究方法两者间并无相互关系。即使研究的功效很低（例如样本量太小，不能适当检测出所研究的效应），仍然能进行严格的测试，如进行临床试验时采用随机化分组。其实，期刊评审指的是你研究的功效并不足以把研究时所观察到的效应当成是可靠和可复制的。 不幸的是，当研究完成后，功效就很难再修改。因此，在开始收集数据前，请先向统计学家进行谘询，确定研究的设计是否有足够的功效。现在这个阶段，你的选项包括把稿件投给一份对功效要求较不严格的期刊，或者进行更深一层的实验以克服此限制。 祝你好运！ ﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎﹎ 英文原文 My paper was rejected because my study was “underpowered”? What does this mean? I was careful about using the best methodology. In statistics, “power” refers to the ability of your study to identify effects of substantial interest. Basically, at the time of designing your study, you need to consider four essential factors: 1. Sample size, i.e., the number of units (e.g., patients), usually represented as “N.” 2. Size of the effect that you are interested in (usually, if you are looking for a large effect, you don’t need as big a sample as you would if you were looking for a small effect) 3. Alpha level: This is your significance threshold (it can be .001, .05, or .1). If your p values are at or above this level, you say that your result is not statistically significant. 4. Power: This is a value representing the likelihood of you finding an effect. How do you determine the power of your study? The above four parameters are interrelated, so if you have the values for three of them, you can calculate the value of the fourth. But usually, the alpha level is fixed (you generally have to choose between .001, .05, and .1) and by reviewing the literature, you will know roughly how large or small your effect can possibly be (effect size). So if you want your study to have good power, you will need to focus on sample size. Many prestigious journals like Nature require you to justify your sample size, so as to show that you have enough power. Nature also offers specific guidelines about what kind of tests you should conduct when your sample size is small. Others, like the British Journal of Surgery , want power calculations to be clearly stated in the manuscript. Still others, like Molecular Genetics and Metabolism , clearly state that “ ubmitted manuscripts without a power calculation will be rejected and returned to authors without review. ” And it’s not just medical and life science journals that are strict about statistical power—the American Psychological Association also strongly recommends reporting a power analysis in the methods section of psychology papers, in its Reporting Standards for Research in Psychology . It also helps to show your power calculations when applying for a grant, so that reviewers can gauge the robustness of your study. You’ll notice that your methodology has not been mentioned in my explanation. This is because your power is independent of your methodology. You can conduct the most rigorous tests, such as randomized clinical trials, even if your study has low statistical power (e.g., your sample size is too small for you to appropriately detect the effects you have chosen to study). What the journal reviewer means is that your study does not have sufficient power for the observed effects to be considered reliable and reproducible. Unfortunately, it’s very difficult to fix power after you have conducted your research. It’s therefore important to consult a statistician before you start data collection, to check whether your study design has enough power. Some of the options you have available at this stage are to choose a journal that is not very strict about power or perhaps to conduct further experiments to overcome this limitation. I wish you much luck. ∷ Eddy 博士国际期刊发表支持中心内容由 意得 辑 英文论文翻译 专家 团队 支持提供 ∷ 【意得辑提供专业 英文论文编校 、 学术论文翻译 、 英文期刊发表一站式服务 www.editage.cn 】 ____________________________________________________________________________________________ 此文同步刊载于 意得辑专家视点 频道： http://www.editage.cn/insights/研究设计中统计功效的重要性
Good study design and forward planning 在本帖中， 理文编辑 学术总监Dr. Daniel McGowan将向大家展示如何做好研究设计和预先规划工作。 Rejection following peer review can mean a considerable amount of additional work for many authors to get their studies published. In the worst cases, their studies may be simply un-publishable. Much heartbreak and hard work can be avoided by simply planning and designing your study properly in advance. In the long run, this will save you time, allowing you to get on with the research for your next big paper. No-one wants to have to repeat experiments because the controls were inappropriate or the case/sample numbers were insufficient to provide enough statistical power. Frequently though, researchers rush into experiments without making all the proper considerations, and this can result in delays when their manuscripts reach the peer review stage. Remembering a few basic principles of study design can help to reduce the risk of outright rejection and repeated experimentation. 1. Have a hypothesis or research question Having a hypothesis or appropriate research question enables you to frame your research within an appropriate context, which in turn will help you apply the appropriate controls. It will also help you describe the rationale for your study when it is time to write it up. Having a hypothesis also means that the objectives of the study are clearly defined, thus reducing the chance that your study will be open-ended and possibly criticised for being incomplete. You can then logically work through these objectives and, importantly, present your results in a logical manner rather than haphazardly. 2. Ensure that the appropriate methods are used Once you have a clear idea of the aims of your study, and the specific research question you are setting out to answer, you will need able to determine what methods would be appropriate to achieve these. Important considerations include deciding whether subjective, qualitative data will be sufficient to address your question, or whether there is a need for more quantitative methods. For basic science studies, such considerations might include the following questions. Will the combination of RT-PCR and in situ data be enough, or is there a need for qPCR? Is Western blotting alone sufficiently sensitive or do you need to also perform immunohistochemistry and cell counting experiments to show a difference between groups? For clinical studies, important considerations include the choice of controls, sample sizes, statistical tests and approach, all of which are described in more detail in the points below. 3. Ensure that the appropriate controls are used Controls are included in experiments to rule out alternative hypotheses. Theres an old saying that nothing can be proven, only disproved, and this is precisely why appropriate controls are necessary: to disprove any feasible alternative interpretations of the data you obtain and/or to eliminate or minimize the effects of extraneous variables. Consider what alternative hypotheses exist, and systematically rule them out by performing experiments that disprove them. There are generally two types of controls: positive and negative. Positive controls show that a negative result is not due to a failure of the experimental system. Negative controls provide an indication of the background noise or baseline value with which to compare values from your experimental sample. In quantitative studies, a relative control or housekeeping control is required to show that changes in the apparent levels of a target gene or protein are not caused by differences in the amounts of protein or DNA in the sample. These levels can be used as a baseline to measure changes in relative levels of a target gene or protein. Common housekeeping molecules include -actin and GAPDH. In clinical trials, subjects in a placebo group in intervention trials, and normal control subjects in observational trials, need to be matched as closely as possible to those in the treatment or disease group in terms of age, sex and numerous other potential confounding factors. In randomized controlled trials, accepted procedures for assignment to groups also need to be followed (see, for example, the ICH good clinical practice guidelines at: http://www.ich.org/LOB/media/MEDIA482.pdf). 4. Use sample sizes large enough to provide a definitive result Many studies fail to achieve the desired impact or to fully support a given hypothesis because the effect is too small or the variability too large to show statistical significance. Often this can be simply overcome by increasing the sample size. However, once a study has been performed and the data analyzed, it can be impossible to go back and increase the numbers without starting all over again. For this reason, pilot studies are often performed in advance of larger scale studies. Talk to a statistician. Determine the size of the effect of your treatment and/or the variability in your population before starting large-scale studies, and use this information to determine the sample size required to give you statistical power. Doing this can save you time, money and potential disappointment later. 5. Use appropriate statistical tests to analyze your data Statistical analysis of your data is essential to show that an effect is genuine and significant. Tests of significance demonstrate the robustness of your findings, essentially showing how unlikely it is that your findings were obtained by chance. Are your data continuous or discrete? Are they normally distributed or non-normally distributed? The nature of your data will determine how they should be analyzed and what tests are appropriate. If in doubt, consult a statistician who will be able to advise you on the most appropriate tests to use and what these tests indicate. Determining the right tests to use in advance will save you having to repeat your analyses if you got it wrong first time round, with the distinct possibility that no significant effect will be observed when the appropriate tests are used. For clinical trials, the following guidelines may be useful: http://www.ich.org/LOB/media/MEDIA485.pdf. 6. Remove investigator and patient bias Many experiments involve subjective measurements or assessments performed by the investigators, as opposed to objective results provided by the experimental system. If the investigator has prior knowledge of the groups to which individuals/samples belong, then investigator bias is a distinct possibility, and this can invalidate any of the findings obtained. In such cases, where the investigator is a factor inherent in the experimental system, it is essential that the investigator is blinded to the groups to which individuals or samples belong. Doing so ensures the objectivity of the findings and improves their reliability. Such blinding can refer to treatment in an intervention trial, or to assessment or interpretation of clinical findings in an observational trial. Similarly, the outcome of a treatment could be influenced if a patient knows if they are receiving a placebo or drug; such patient bias should be avoided, by blinding the patient to the nature of the treatment. Being aware of the potential for bias before commencing experimentation can again save the need for time- and resource-consuming repeats. 7. Comply with ethical requirements There are strict regulations regarding the use of human and animal subjects, and in many countries, regarding the use of stem cells, cell lines and genetically modified materials. Failure to comply with these regulations will prevent publication of your findings and could lead to legal issues; at best, it will limit the range of journals to which you can submit your findings. Make yourself aware of these regulations before you commence your study and ensure that all requirements are complied with so you dont encounter problems later on. As well as ethical requirements regarding experimentation, there are also strict guidelines provided by most journals regarding the requirements for authorship, and these also need to be complied with. Clinical trials should comply with the Declaration of Helsinki (http://www.wma.net/e/policy/b3.htm) in addition to any local requirements. Informed consent is essential for most trials involving human subjects. Animal studies should comply with local and national regulations, although many journals are now aligning themselves with standards such as the NIH Guidelines for the Care and Use of Animals (http://oacu.od.nih.gov/regs/guide/guide.pdf). Finally, many journals require a statement describing who gave ethical approval for the study. 8. Clinical study registration Many top-tier journals now request that prospective clinical trials involving human participants should be registered online in an accessible database. Many journals will instantly reject studies of this type that have not been registered. More information on this can be found at http://www.icmje.org/faq.pdf. International clinical trial registries include the Chinese Clinical Trials Register (http://www.chictr.org/), the Japanese Primary Registries Network (http://rctportal.niph.go.jp/), The International Standard Randomised Control Trial Number database (http://isrctn.org/) and Clinical Trials.gov (http://www.clinicaltrials.gov/). Registration should be done before the first participant is enrolled, but many of the databases do allow retrospective registration. However, by registering the trial once you receive ethical consent you will save time and overcome a major obstacle to publication. All studies are different and therefore have different requirements regarding appropriate study design. The points above are just a few of the important considerations that should be made prior to the commencement of experimentation, and the general principles apply to a variety of different study types. It is true that sometimes even peer review fails to detect flaws in study design, as shown, for example, in the following report on randomized controlled clinical trials published in Chinese journals: http://www.trialsjournal.com/content/10/1/46. However, if you want your study to stand the test of time, be published in a top-tier journal and to be widely accepted by the international research community, then planning ahead and designing your study to make it robust and reliable will only serve to save you time, money and heartbreak later on. 在这里还需提请各位注意，Dr. McGowan 的母语是英语，无法阅读中文，因此请大家尽量使用英文回帖，如有任何需要与他沟通的学术和语言问题也请使用英语，Dr. McGowan 会及时回复大家的。 Dr. Daniel McGowan 曾任 Nature Reviews Neuroscience 副编辑，负责约稿，管理和撰写期刊内容。于2006年加入理文编辑（Edanz Group） 并从2008年起担任学术总监。Dr. Daniel McGowan 有超过十年的博士后和研究生阶段实验室研究经验，主要致力于神经退化疾病、分子及细胞生物学、蛋白质生物化学、蛋白质组学和基因组学。
标签: 研究设计
