Main content



Loading wiki pages...

Wiki Version:
- Proposed analysis pipeline, including all preprocessing steps, and a precise description of all planned analyses, including appropriate correction for multiple comparisons. Any covariates or regressors must be stated. Where analysis decisions are contingent on the outcome of prior analyses, these contingencies must be specified and duly followed. Planned analyses can be restricted to a well-written and well-commented script and do not have to be described separately in the Wiki of the component. Only pre-planned analyses can be reported in the main Results section of Stage 3 submissions. However, unplanned exploratory analyses will be welcome in a separate section of the Results called Exploratory Results (see below). - Studies involving Null Hypothesis Significance Testing (NHST) must include a statistical power analysis. Estimated effect sizes could be justified with reference to the existing literature. However, since publication bias is known to inflate published estimates of effect size, it is usually more desirable to base one’s power analysis on the theoretically smallest meaningful estimate of the effect size (see Perugini et al. [2018] for helpful pointers). For frequentist analysis plans, the a priori power must be 0.8 or higher for all proposed hypothesis tests. One-tailed tests, if properly justified by the hypothesis, are encouraged. In the case of highly uncertain effect sizes, the use of a variable sample size and interim data analysis is permissible but with inspection points stated in advance, appropriate Type I error correction for ‘peeking’ employed, and a final stopping rule for data collection outlined. In cases where large samples are practically impossible, authors should contact the editors beforehand outlining their strategy. - Methods involving Bayesian hypothesis testing are encouraged. For studies involving analyses with Bayes factors, the predictions of the theory must be specified so that a Bayes factor can be calculated. Authors should indicate what distribution will be used to represent the predictions of the theory and how its parameters will be specified. For example, will you use a uniform up to some specified maximum, or a normal/half-normal to represent a likely effect size, or a JZS/Cauchy with a specified scaling constant. For inference by Bayes factors, authors must be able to guarantee data collection until the data is at least 6 times more likely under the alternative hypothesis as opposed to the null hypothesis (or vice versa). Authors with resource limitations are permitted to specify a maximum feasible sample size at which data collection must cease regardless of the Bayes factor; however, to be eligible for advance acceptance, this number must be sufficiently large that inconclusive results at this sample size would nevertheless be an important message for the field. For further advice on Bayes factors or Bayesian sampling methods, prospective authors are encouraged to read this key article by Schönbrodt and Wagenmakers. - If described in the Wiki, any description of prospective methods or analysis plans should be written in future tense.
OSF does not support the use of Internet Explorer. For optimal performance, please switch to another browser.
This website relies on cookies to help provide a better user experience. By clicking Accept or continuing to use the site, you agree. For more information, see our Privacy Policy and information on cookie use.

Start managing your projects on the OSF today.

Free and easy to use, the Open Science Framework supports the entire research lifecycle: planning, execution, reporting, archiving, and discovery.