Welcome! This repository is a place for researchers to post analysis plans for projects using data collected with the Kluane Red Squirrel Project.
Analysis plans are distinguished from project proposals in that they consist solely in a detailed outline of the data analysis a researcher intends to carry out (i.e., no need for literature review or theoretical background). For studies using the long-term dataset, this should include details on which data will be extracted (e.g., grids, years, any exclusions). Planned statistical procedures should then be presented as unambiguously as possible. Where analysis decisions need to be informed by data, the researcher can present “what if” scenarios with decision criteria (e.g., if we detect overdispersion, we will fit X model).
This should serve as an informal form of preregistration, with the aim of helping researchers to avoid p-hacking or hypothesizing after results are known (“HARKing”), even if inadvertent. Accordingly, plans should be posted before running any analyses. However, having a posted analysis plan does not restrict the researcher from subsequently adding additional or follow-up analyses. Best practice would require that should be clearly delineated as additions to the initial analysis plan in any published works.
Lastly, it is worth noting that the intended benefits of preregistration mostly apply to research setting out to test hypotheses. For exploratory projects designed to create new hypotheses without providing robust inferences, pre-posting plans may not be helpful.
More resources on preregistration can be found at:
- https://www.cos.io/initiatives/prereg
- Nosek, B. A., C. R. Ebersole, A. C. DeHaven, and D. T. Mellor. 2018. The preregistration revolution. Proc. Natl. Acad. Sci. U. S. A. 115:2600–2606. www.pnas.org/cgi/doi/10.1073/pnas.1708274114