-
-
Notifications
You must be signed in to change notification settings - Fork 151
Sealing or redacting a document on CourtListener
In general, our policy is to never alter court materials or remove anything from CourtListener. That said, we balance this against court orders that seal or redact content, always trying to find a happy medium between our goals of transparency and permanence vs. litigants important needs for privacy. It's not always an easy line to walk, but most cases are straightforward.
Follows is how we approach an inbound request to remove or redact something.
First, decide if it's from a case of national interest. If the removal is of a newsworthy document, we might fight it. That hasn't happened yet, but it's worth thinking about.
Second, decide if the person looks like an attorney.
Respond to the request. We need:
- A link to the item. (This ensures we're all talking about the same item and generates an unambiguous paper trail.)
- An attached order to remove the item. (This ensures we only get legit requests.)
We need enough information to properly identify the item. If you have to write back to the requestor about that, you might as well ask for the sealing or redacting order while you're at it, since sealing orders are often slower than attorneys that want something taken down. The attorney may have it, before the clerk has put it on the public docket.
Begin reviewing it. Questions you should ask yourself:
-
Is the sealing/redacting order from the same court as the item they want removed?
- If not, the order is invalid. Higher/lower court orders do not apply.
-
Is the item still available on the court website?
- If it's a state court website and the item is no longer available, that can mean anything. State courts remove stuff all the time.
- But, if it has been removed from PACER, that's unambiguous. We can remove it from our system too. Great.
-
Is the sealing/redacting order available on the docket?
- If so, great.
- If not, we either:
- need it from an attorney or
- we cannot proceed.
So at this point, you should have a valid order from the correct court or the item is removed from PACER. Great. Let's remove it.
PACER has three levels of sealing/redacting. Items are never modified in place that I'm aware of. The three levels are:
-
The docket is sealed.
This is the most extreme case and warrants the most careful review. It's rare, and it takes work. If it's a recent docket with no opinions, you can probably remove the whole docket. If it's an older case that has only recently been sealed, we might fight that.
If we are going to remove it, then we'll use the Django admin panel to do so.
First, follow the procedure below to remove all the documents on the docket. Then, use the admin panel to delete the docket itself. As you do this, Django will tell you if anybody has any interaction with it, such as favorites, tags, or alerts. If so, consider the pros/cons of notifying those users before/after you delete the docket. There's probably a good argument that they should be notified of the removal.
With the users notified (or not), delete it from the database and notify the requestor. Done.
-
The docket entry is sealed.
This happens occasionally. To fix this, first, redact the associated document as below. Then delete the docket entry by using the admin panel.
-
The document is sealed.
When you delete a database record from Django, it doesn't delete associated PDFs, and it certainly doesn't delete things from Internet Archive. Therefore, to delete sealed items from CourtListener, you have to use a special approach. To do this, first grab the RECAP Document ID. You can find this in the links in the sidebar of a document page:
For example, the link might be:
https://www.courtlistener.com/admin/search/recapdocument/352873075/change/
Which would have the ID of:
352873075
Using that ID, make a link like:
https://www.courtlistener.com/admin/search/recapdocument/?q=352873075
You'll see something like this:
Check the box next to the item on the page, and use the Actions drop down to "Seal Document"
That will delete the item from Internet Archive and storage.courtlistener.com, and will remove any data we gleaned from the document (like its page count, hash, text, etc.)
These are usually easier. Just delete the items from the databaes, and as above, if there are any user-related content associated with them, make a decision about whether to notify those users.
This comes up from time to time.
Usually you can just upload the latest content using RECAP and you'll get the fix.
You may have to do a little manual surgery to the HTML we have in our system. Just use CTRL+F to find all the instances that need to be redacted, and do so by hand using common sense.
If the PDF/DOCX/DOC/WPD/etc file needs modifying, the best solution is to find the updated document from the court. It's often at the URL of the original item, but if you click the original item and it still brings up the unredacted version, that's often a hint that we don't have to do more. If the court hasn't updated their copy, that's fishy.