Hello, tumblr! I saw something on here the other day that worried me, so I decided to Do Science about it. But I can't do it alone: I need your help to build the dataset!
Here's what I need you to do:
If you see a post with a "mature content" label, and it's 2026, DM me a link to the post.
Yes, that's really it.
I am hoping to collect several thousand such posts, so that I have a decent sized dataset. I do not care what the post is about; if it's labeled as "mature content", I want to add it to my dataset.
If I get 10,000 posts in my dataset before August 31st 2026, I will post my preliminary findings then. I won't feel comfortable calling my findings "settled" before 2027, unless I get over 50,000 posts.
Tumblr Science FAQ, round 1
What's your hypothesis, OP?
I am not talking about that unless I have results to share. That would bias the results.
I did write them down and I did share them with a trusted contact who can prove that I wrote them down the same day I made this post. (While I did so before I made this post, I am not sure they will be able to provide proof of that, because I did so on the same day.)
OP are you interested in...?
Do you have to click through to see the post? Does the clickthrough contain the words "mature content"? Then yes.
OP are you interested if the post is about...?
I am interested in the mature content labels, not the content of the post. Is there a clickthrough that contains the words "mature content"? Then yes.
Tumblr Science FAQ, round 2
Should I reblog this for reach?
Yes, please. I felt really silly when I noticed I forgot to include that in the original post.
Is it okay if I send you my own posts?
Yes, those are perfect for what I'm looking for. I actually need to do some processing on a post to make it useful for testing my hypothesis, and this makes it easy.
Are you looking for "potentially mature content" also, or just "mature content"?
I want both, please. Anything that throws up a blocking screen that you have to click through. The distinctions between them are one of the things I am hoping to study.
Does it matter when the post was made originally?
Technically no. There's no way to respond to this without introducing some bias in the results, and I don't want that. However, I do collect some data on a post as part of making it useful to me, and that data is easier to collect if the post is recent.
What if I request content label review on a post after I send it to you?
I need to see the mature content label to be able to use the post. Because the mature content label hides the content of the post, it is very hard to use a post that no longer has a mature content label. You could send me a screenshot, but people could use that to lie to me.
Basically, it's more work for you to make it usable to me.
OP are you a transphobe? It would ruin the experiment for me if you're a transphobe.
I promise I'm not a transphobe and not doing this for transphobic reasons. You should still double-check that I'm not a transphobe for yourself, though. I am not sure that this study will have the useful effects you're hoping for; I am studying something specific, and it may not be what you hope.
Tumblr Science FAQ, round 3
posts created during 2026, or posts shared with you during 2026?
Posts shared with me during 2026.
It is easier to use a post if it was made in 2026, but it doesn't have to be made in 2026.
Some posts are labeled mature content by Tumblr, and some posts are labeled mature content by the person making the post. Won't that change the result?
That's actuality what I'm studying!
Is there a way I can follow updates, without having to follow you?
As of this FAQ post, I am using "iamthelowercase science project updates" as a tag on all my posts about it. I am considering whether to use tumblr's "community" feature to create a dedicated place to follow about it, as well.
If I'd thought about it beforehand, I'd have made a sideblog specifically for this project. Hindsight really is 20/20.
Tumblr Science FAQ, round 4
OP are you secretly rounding up a bunch of posts to report them and get them deleted?
This question comes in two parts. Part one: well if I were doing it secretly, I wouldn't tell you that, now would I? I promise I'm not doing that, but like, isn't that what I would say if I were lying to you?
Part two: LMAO if you think Tumblr doesn't already have the ability to nuke all posts marked "Mature Content" just as thoroughly as they nuke trans women's blogs. Now like, a mass reporting campaign targeted at this dataset could still have negative effects, if nothing else by getting employee eyes on the posts. Which is why I'm going to have to be careful who I share the full dataset with. But that's antithetical to my goal here.
What if I have tumblr set to not show me the mature content clickthrough? Do I still send you the post if I notice it?
Yes please. I am studying the clickthroughs themselves, more than the content of the post; but for the same reason I am looking at the clickthrough myself. You don't have to look at the clickthorough if you don't want to.
I am specifically comparing posts labeled "potentially mature content" to those labeled "mature content". Some people have turned the clickthrough filter back on so that they make sure to notice when posts they would not consider mature got flagged. I am very grateful to these people.
What if the post is from a few years ago?
Go ahead and send it. Some of those will prove useful to me, and some of them won't, but the only way for me to know for sure will be to check them.
I just didn't want to be getting posts from this project forever.
Things I've learned so far
I should have done a trial run before blazing the post. Oops.
Apparently there's some sort of scam where people need to trick you into messaging them first? And this does look like it would be one of those. I promise it's not one, but of course you have no way to know.
You can just. Tell people the null hypothesis without having to worry about it corrupting your data.
Just how much work data entry is.
I should have provided some way for folks who are interested in this project to receive updates on it without resorting to following me. Sorry if you followed me for information about this project and were disappointed in my regular posting.
Whether or not a pinned post can be a reblog (this is the test case!).
You can paste markdown into the "create post" form on web, and the formatting gets automatically applied. Handy.
That doesn't work on the Android app, though.
Tumblr uses github to track past versions of their community guidelines (good! +1 trust), but they didn't do so when they last updated the community guidelines (bad! minus a whole bunch of trust).



















