When "likers'' go private: Engagement with reputationally risky content on X

The finding is surprising, but I think their methodology is a bit flawed.

Study 1 shows "Difference-in-Differences analysis of engagement with 154,122 posts by 1068 accounts before and after the policy change". All this tells us is that existing accounts did not have a noticeable change. It doesn't suggest anything about accounts created after where the culture of Twitter (appears) to have shifted quite a bit from before going private.

Basically "okay cool, existing accounts didn't change their behavior". What about new accounts? More anonymous accounts? Can we understand anything else about platform growth and interaction? What about classes of user w/ respect to verified users, anonymous accounts vs accounts tied to real identities?

Study 2 is also very limited to draw that conclusion because people are less likely to honestly report their engagement with content or beliefs that could be punishing in a given political environment. This was most astutely observed by the French polymarket user who crushed it betting on the 2024 election using neighbor-polling methodology [0]. Essentially, it appears to be more reliable to ask about the preferences of a respondent's social circle than ask the respondent directly.

[0] - https://www.cbsnews.com/news/french-whale-made-over-80-milli...

If it can be set to private then it can be set to public again. I don't use any of those platforms but I would always assume that all my usage might end up being published one day.

The vast majority of "likers" have never been real people in any case. All of the prominent accounts are boosted by bots and Mechanical Turk users in economically underdeveloped countries. This has been shown numerous times by comparing the likes/impressions ratios for different accounts posting similar content.

Anecdotally, I have been 'liking' (as a verb) posts about 3x more after anonymity went into effect. I used to be anonymous on X until I started meeting people at IRL events and then had to be more cautious about what I broadcast to my network. Anonymized likes gave me back a lot of that freedom.

I can't find any mention of paid versus free accounts in this study. It used to be the case that people who paid for Twitter were already able to hide their "likes", before Twitter just made "likes" hidden for everyone. I would be interested in knowing if the visibility change caused anyone to give up their subscriber status, i.e. those people who would pay extra because they really care about keeping their "likes" hidden.

Note that Twitter "likes" is still not private today in the sense that the original post authors can see who liked their post. I suspect people who were really sensitive to this visibility simply wouldn't engage with risky content to begin with.

I'd say this study is inherently flawed. As I am sure most people know on the Internet these days that just because X states their 'likes' are 'anonymous', doesn't mean they are.

I think the potential reputational damages would still be on the forefront of most people's minds, knowing that at any stage, at the whim of Elon, these will be revealed.

> We find no detectable platform-level increase in likes for high-reputational-risk content (Study 1). This finding is robust for both between-group comparison of high- versus low-reputational-risk accounts and within-group comparison across engagement types (i.e., likes vs. reposts). Additionally, while participants in the survey experiment report modest increases in willingness to like high-reputational-risk content under private versus public visibility, these increases do not lead to significant changes in the group-level average likelihood of liking posts (Study 2).

That conclusion's a surprise to me. I used to basically never like anything (even innocuous stuff) unless I specifically wanted to endorse it (essentially treating it as a less direct retweet). I like stuff all the time now.

They do note their methodology could be affected by inorganic engagement that wouldn't be affected by like visibility, though. I wonder what other factors could've led to that conclusion.

A lot of people are going to be upset by the idea of their likes being public, but I really like and hope we see better analysis of likes on Bluesky/atproto, where this data is public!

Imo it really sucks they social networking is a dark forest, controlled by a very few, who increasingly have offered less and less and less at higher and higher prices to researchers, academics, and more generally bots and services that used to be up to & doing cool things. BlueSky has the juice, imo, and while most folks using it today are only using official Bluesky services, some folks are using independent services for all their PDS hosting and for viewing the network.

That the network is public feels like such a minimum baseline level, is such a basic obvious and essential baseline for society to begin to have any trust ability or engagement with such mass communication systems as we have.

What's the difference a private like and a bookmark? What's the difference between a public like and an RT? They can be tracked separately, but is that necessary?