Information for Journal Editors
RIDIE offers a means to ensure a research project's credibility, lack of bias, and quality of impact. In particular, prospective registration of a study — under which evaluators lay out hypotheses, subgroups for analysis, and other design details — minimizes the potential for post hoc data mining or specification searches, whereby researchers use the results to decide what outcomes to report or specifications to use.
In clinical medical research, registration and journal publication are tightly linked. Since 2005, the International Committee of Medical Journal Editors (ICMJE) has had a policy requiring investigators to deposit information about their randomized controlled trial designs into a clinical trial registry (such as ClinicalTrials.gov) prior to enrolling study participants as a precondition for publication of the trial’s findings in member journals.
While this approach may be too strict to apply to the range of evaluations carried out in the social sciences, journals can still use RIDIE or other registries to ensure the integrity of the evaluation-related research they publish. A number of economics journals already require that authors make available the data set and/or programs they used in their analysis. Giving preference to impact evaluation-related research that made use of RIDIE or another registry to pre-register the evaluation — or making this a requirement of publication — can help to ensure the integrity of published impact evaluations.
Editors may, like ICMJE, specify the elements or fields in the registry that they want to see entered, and they may want to reference the RIDIE registration number in published articles. Finally, editors may want to provide information from the registry (including initial analysis plans and changes to study design) to referees to assist them in evaluating submissions.
The journal can therefore ask authors who submit evaluation studies to indicate if they registered the study at RIDIE (or elsewhere) and provide the registration number to allow easy reference, or alternatively to submit the registry information report directly to the journal. This includes the study's full entry and any associated change history, as well as the study completion data. Upon publication of the paper, the journal can add the RIDIE identifier to the paper's footnote. Interested readers can use this number to access the record in RIDIE just as they can download the data files and program code from the journal's website.