gordpennycook.bsky.social
Associate Professor, Psychology @cornelluniversity.bsky.social. Researching thinking & reasoning, misinformation, social media, AI, belief, metacognition, B.S., and various other keywords. 🇨🇦
https://gordonpennycook.com/
275 posts
22,398 followers
1,127 following
Prolific Poster
Conversation Starter
comment in response to
post
comment in response to
post
There are plenty with *some* evidence of effectiveness, sure: www.nature.com/articles/s41...
Fact-based approaches (prebunking/debunking) are broadly effective. If the outcome of interest is social media sharing, accuracy prompts have shown some effectiveness in real-world tests.
comment in response to
post
Sounds great!
comment in response to
post
Also, big thanks to @profsanderlinden.bsky.social for offering advice at an earlier stage of this research! We ran Study 5 on his advice and it was a great conclusion to the paper.
comment in response to
post
This research was led by @ninawang.bsky.social & Samantha Phillips (both amazing)
It was made possible b/c of funding from the MINERVA program. The grant has since been cancelled, alongside the entire program - thanks Trump! Why bother testing the efficacy of scaled interventions, right?!?
comment in response to
post
Coincidentally, several of the main proponents of inoculation recently posted a manuscript that reports an attempt to test the intervention using ads on Twitter. It was not effective. (major props to the authors for writing this up!) scholar.google.ca/scholar_url?...
comment in response to
post
We conclude, therefore, that psychological inoculation is unlikely to be effective at scale. Social media is a highly noisy environment and the chances that people would be able to spontaneously apply the knowledge they gain from a short inoculation video in a noisy feed is (in my view) low.
comment in response to
post
The conclusion is fairly straightforward: The inoculation video only had a detectable effect in the case where there was no additional distracting information in the feed. If the task was straightforward, the participants appeared to be responsive to the intervention. Otherwise, there was no effect.
comment in response to
post
Study 5 was unique because we *only* populated the social media feed with the (stylized) "synthetic" emotional language tweets from previous work. Thus, this was the scenario that was the closest replication of past work... but, naturally, it is also the most unrealistic scenario.
comment in response to
post
We did several variations on the feed length and we also measured three "real-world" type outcomes: Dwell time, likes, and shares.
The results are pretty straightforward: There were no consistent effects for any outcome. With one exception: sharing in Study 5.
comment in response to
post
1) The inoculation videos were presented as a pop-up in a "mock" social media feed (via "Yourfeed" see arxiv.org/abs/2207.07478)
2) We used both real-world content and the "synthetic" materials that are typical of past research on inoculation (good internal validity, but may not be that realistic).
comment in response to
post
There are two key methodological details that make this (we think) a stronger test of whether psychological inoculation is likely to be effective if deployed on social media at scale:
comment in response to
post
In these experiments, we were testing a previously validated emotional language inoculation video. The goal is to increase people resistance to emotionally manipulative language.
(Side note: I'm skeptical about this approach, but that's a separate argument shorensteincenter.org/new-event/ps...)
comment in response to
post
Oh wow, okay - thanks for the info! The point still stands, of course, but this also serves as yet another example of missing context on social media posts.
comment in response to
post
Thus concludes today's musings.
comment in response to
post
Of course, the (perhaps more plausible) alternative is that they're just lying. They don't really care that they are going to be spectacularly wrong because they can just use their propaganda to distract people from that fact.
Probably both are true. Some are liars & some are fools. Some are both.
comment in response to
post
If you're curious, here's the retracted article: www.pnas.org/doi/10.1073/...
Major props to Duckworth et al. for this retraction
comment in response to
post
Thanks! Yah that would be great!