Frequently Asked Questions

1. We need to change something (e.g., wording, typo, etc.) in our pre-registration, can you change it for us?
We never make any changes to pre-registrations. Authors have the following options to make edits or flag problems:

If it has not been approved yet by all authors, they may reject it, and create a new pre-registration, the answers of the rejected pre-registration can be automatically copy pasted to the new form, and then edited before submitting.

If it has been approved by all authors, and it has been only a few days since its creation, you may start a new pre-registration. You will be asked how it relates to recently created pre-registrations and you may deem the existing one a 'draft', this will lead the new pre-registration to replace the old one, the draft, which will then be deleted.

If it has been more than a few days since the to-be-edited pre-registration was created, or if you have already generated a .pdf, you can no longer delete it or edit it. You can explain in your paper any shortcomings with the pre-registration that you want to alert readers to.

2. I created a .pdf of a pre-registration by mistake, can you delete it?
No. For transparency and credibility of the pre-registration system, once a link to the pre-registration is created, the link must always be available.
At this point in time, we do not offer search and do not index our pre-registrations thus as long as you do not share the link, nobody should come across it. We may change this in the future and provide centralized access to all pre-registrations, but do not currently plan to.

3. I indicated two pre-registrations are related and they were bundled but I made a mistake, they are not related, can you unbundle them for me?
No. For transparency and credibility of the pre-registration system, once something is included in a shared pre-registration it must always be included, and that includes links to other pre-registrations the authors indicated were related.

4. Something is wrong with how my pre-registration is displayed, can you fix it?
Yes. We will fix anything that is our responsibility. Examples include titles that do not fit the .pdf width, and characters that do not display correctly in the pdf.

5. I no longer have access to my old email account, can you help me recover my old-pre-registrations?
Yes. Create a new account with an email address you do have access to, visit the account page and enter the email address of your previous account as a secondary email for the new one.

6. It's something else
For anything not listed above please email us at support@aspredicted.org

Preventing Abuse of Pre-Registrations at AsPredicted

A concern some researchers have is that an author at AsPredicted could create multiple parallel pre-registrations, alternative sets analytical decisions, try them all, and then make public the one that ended up "working". Effectively circumventing the restrictions pre-registration places on p-hacking. This kind of abuse is actually very difficult to do at AsPredicted, as we have multiple mechanisms in place to prevent it.

First, authors, when creating a new pre-registration, need to indicate how it is related to all their recent pre-registrations, including whether the study would share even a single data point with a previously pre-registered study.

If the author says 'yes', the pre-registrations become bundled, and making one public makes both public, and each pre-registration links to all pre-registrations in the bundle. So right off the bat, to do the kind of gaming people fear, authors would need to actively lie. 

If the author says 'no', our algorithms compare the pre-registration content. If they are deemed similar, authors are asked to document separate data collection efforts (e.g., MTurk hits), and if they do not provide those within six months, the pre-registrations are bundled anyway.

Finally, if separate authors submit pre-registrations that are deemed similar by the algorithms, and someone at AsPredicted consider that to have been an effort to circumvent our algorithms (to bypass the question on whether the efforts are related), the accounts can be blocked and the more recent pre-registration be deleted.  This rarely occurs, a handful or fewer times per-year.

When is a Pre-Registration Private vs Public?

Private pre-registrations and shareable PDFs.
A central aspect of a pre-registration platform is how authors share pre-registrations with readers.  At AsPredicted, when a pre-registration is created, only its authors have access to it, there is no way for third parties to access that pre-registration, no link to it exists.  When authors decide to share a pre-registration, for example, in order to link to it from a paper they are submitting for peer-review, one of the authors needs to generate a PDF. It will be anonymous (will not list authors), and will have a permanent URL which will look like this: https://aspredicted.org/b3d4-e41h.pdf

When the paper is accepted, the PDF should be updated so that it is no longer anonymous. Importantly, when that happens, the URL will not change. The URL that is first generated for a pre-registration will always be its only URL, and only the target of the link, the PDF file, will be updated to be anonymous or not.

Reversibility 
The act of creating a PDF is non-reversible. Once a PDF exist, once a URL linking to a pre-registration is created, it cannot be deleted, even if created in error.
In contrast, the act of making a PDF anonymous or not-anonymous, is reversible. At any point authors can re-generate the PDF so that it does or does not reveal author names. When they do, the URL remains unchanged.

Policy up to September 2024.
From nearly its beginning in 2015, until September 2024, AsPredicted PDFs were handled in a slightly different way. Pre-registrations used to have three modes, "private", "reviewer", and "public". Private mode has not changed.  When authors created a pre-registration only they had access to it, there was no external link to it. When authors wished to submit to a journal an anonymous version of the pre-registration, they created a PDF in "reviewer" mode, which was a PDF that was anonymous and indicated visibly that it was only intended for peer-review. Early on the links to those PDFs looked like this,: https://aspredicted.org/blind.php?x=asdsje, and then they looked like this: https://aspredicted.org/ABC_DEF. 

Pre-registrations in reviewer mode were not backed up to the WebArchive (because they were supposed to be temporary).

The third mode was "public", which involved a permanent PDF that listed the authors, with URLs like this: https://aspredicted.org/abcde.pdf
These PDFs were backed up to the WebArchive.

Problem with the old policy
The reason we changed this policy is that authors often forgot to update the AsPredicted links in their papers when they got accepted, and thus many published papers link to the anonymous and not backed-up version of the pre-registrations. In September of 2024 we made a forward looking and a backwards looking change. The previous section explains the forward looking one, the new policy. In terms of the latter, we asked authors who had any anonymous PDFs created already, whether they would consent to having those be backed-up on the WebArchive if they were found in the public domain, which we operationalize as having been found by a major search engine crawler like Google or Bing. Because these engines crawl academic papers, pre-prints, and public websites, but do not crawl personal files or emails, these crawler provide an efficient way to locate anonymous pre-registrations that made their way to the public record and thus should be backed up. We will back up a pre-registration if any of its authors consents to the new policy.

Pre-registrations PDFs are permanent and backed up to the Web Archive.

All pre-registrations PDFs at AsPredicted are permanent, authors may not delete them or hide them after they are created.

In addition, PDF files are backed up to the WebArchive. The back up to each AsPredicted URL is simply the original URL plus the following text before it: http://web.archive.org/web/*/.

For example, the backup for the pre-registration available from
https://aspredicted.org/kv692.pdf, is found at:
http://web.archive.org/web/*/https://aspredicted.org/kv692.pdf

PDFs are backed up when any of the following events occur:

1. An author creates a non-anonymous PDF for their pre-registration
2. A web crawler for a major search engine (e.g., Google and Bing) visits the URL for the anonymous PDF, indicating a link to such pre-registration has been found on a public document.
3. Two years elapse since the creation of the anonymous PDF

The triggers described in (2) and (3) were added in September of 2024. They are applied to anonymous PDFs created before such date if at least one of the authors of the pre-registration agreed to the policy change.

Policies for anonymous PDFs.

They:

• Do not show authors' names.
• Should be deanonymized by authors when their paper gets accepted for publication.
• Have a unique and permanent URL that will not change when the PDF is deanonymized.
• Will be backed up to the WebArchive (archive.org) when any of the following happens:
  • An author deanonymizes a pre-registration
  • We discover a link to the PDF from a public document (e.g., pre-print or journal article)
  • Two years pass after the PDF is originally created
• Are automatically deanonymized after 4 years (authors can reanonymize)
• Are only discoverable if authors share their URL.
  But: AsPredicted may make URLs for non-anonymous PDFs discoverable in the future (e.g., via search or open access).
• Cannot be deleted or edited, their contents and URL are permanent.