Current Events > Study: little evidence that Russian bot campaign actually influenced humans

(+) Yeah! (+)

https://www.nature.com/articles/s41467-022-35576-9

There is widespread concern that foreign actors are using social media to interfere in elections worldwide. Yet data have been unavailable to investigate links between exposure to foreign influence campaigns and political behavior. Using longitudinal survey data from US respondents linked to their Twitter feeds, we quantify the relationship between exposure to the Russian foreign influence campaign and attitudes and voting behavior in the 2016 US election. We demonstrate, first, that exposure to Russian disinformation accounts was heavily concentrated: only 1% of users accounted for 70% of exposures. Second, exposure was concentrated among users who strongly identified as Republicans. Third, exposure to the Russian influence campaign was eclipsed by content from domestic news media and politicians. Finally, we find no evidence of a meaningful relationship between exposure to the Russian foreign influence campaign and changes in attitudes, polarization, or voting behavior. The results have implications for understanding the limits of election interference campaigns on social media.
...
Here, we demonstrate the extent to which exposure to posts from foreign influence campaigns was concentrated among a small group of users. We show, further, that although the amount of exposure to social media posts from foreign influence campaigns is large in absolute terms, it was overshadowedby at least an order of magnitudeby content from ordinary domestic political news media and US political candidates. Finally, we examine whether exposure to posts from the Russian foreign influence campaign is associated with changes in US respondents positions on salient election issues, levels of political polarization, and voting behavior in the 2016 election. Across a wide range of outcomes, we do not detect a meaningful relationship between exposure to posts from Russian foreign influence accounts and changes in attitudes, polarization, or voting behavior.

Wouldnt the whole point be to make it not easily traceable?

and its not like people are going to admit they got influenced by Russian bots

I mean it's kinda hard to demonstrate what influenced people

no right wing chud is gonna be like "Man these Russian bots really speak to me"

But being reaffirmed of all your worldviews by all media you consume over the course of years will effect you

thats not going to stop liberals on twitter acting like they are in a spy movie being assailed by kremlin spies on all sides

Lmao of course people here are in denial about these results due to confirmation bias. Stop ignoring evidence because it doesn't support your theories.

Weird to assume bad foriegn actors have no influence on domestic news, particularly when some domestic news agencies tend to blast the airwaves with bullshit like that, such a OAN.

UnfairRepresent posted...

no right wing chud is gonna be like "Man these Russian bots really speak to me"

Luckily that is not how they attempted to study this issue.

To investigate the relationship between exposure to Russian foreign influence accounts and political attitudes we take advantage of the panel structure of the data. These data allow us to examine within-subject variation, and therefore to account for time-invariant characteristics across respondents. We model the relationship between exposure and issue positions and ideology by regressing within-respondent changes in issue positions and perceptions of polarization on exposure to posts from Russian foreign influence accounts between survey waves. In other words, we examine whether exposure is associated with changes in each respondents political attitudes and perceived polarization from before the campaign to immediately prior to the election. If exposure were unconfounded between survey waves (a strong assumption), the estimand in this model would be the average treatment effect on the treated to the extent that levels of exposure is as observed in the data. The regression model is also equivalent to one that predicts a respondents attitude toward an issue in the final wave of the survey conditional on the position that they took on that issue in the first wave of the survey (with a coefficient of 1) and their exposure to posts from Russian foreign influence accounts. Because exposure to foreign influence accounts is concentrated among a relatively small group of users (as shown in panel b of Fig. 1), we note that our estimates in the following sections are driven by those (heavily) exposed to messages from these accounts. Were exposure distributed differently, among another set of users, the estimated relationship could well be different. It should thus be kept in mind that those exposed to foreign influence accounts were users who self-identified as highly partisan Republicans, a fact that in itself aids in contextualizing the limited scope of the Russian foreign influence campaign.

The main results are presented in Fig. 4 (see Supplementary Methods F for complete output). As the figure shows, we do not find statistical evidence in support of a relationship between exposure to posts from Russian foreign influence accounts and changes in respondents issue positions or perceptions of polarization. This is the case regardless of whether exposure is measured as the (log exposures + 1) number of exposures to posts from Russian foreign influence accounts or as a binary variable indicating whether a respondent was exposed to at least one such post. For ideological and issue positions, the estimated relationships are neither significant nor in a direction consistent with one favorable toward Donald Trump. This result is similar for polarization. Of the two statistically significant coefficients across all models (representing 6% of our coefficients, as would be expected by chance), neither is in a direction that would suggest that exposure to posts from Russian foreign influence accounts is related to an increase in perceived polarization. Finally, adjusting for multiple comparisons (see Supplementary Methods G), we do not find statistically significant relationships between exposure on any of the issue-based and polarization outcomes.

Regardless of how effective they were, we all agree we shouldn't ignore any attempts to subvert our democracy, right?

[LFAQs-redacted-quote]

Going further, any such study that doesn't include Fox and the like in the same brush as Russian bots is inherently bunk in nature. Just because it is a real person does not change the source of the issue.

COVxy posted...

Weird to assume bad foriegn actors have no influence on domestic news, particularly when some domestic news agencies tend to blast the airwaves with bullshit like that, such a OAN.

To be clear, I think the primary issue with this line of work is that it only tries to examine direct contacts. Not only do domestic news agencies pick up on these "stories", but the people that do get influenced by direct contact are probably much more likely to influence other people slightly less far out on the spectrum than them, and then those people are more likely to influence people slightly further out, and so on and so forth.

Seems obviously flawed to look at a complicated network and only look at the effect of direct influence.

[LFAQs-redacted-quote]

what in the butt.

COVxy posted...

To be clear, I think the primary issue with this line of work is that it only tries to examine direct contacts. Not only do domestic news agencies pick up on these "stories", but the people that do get influenced by direct contact are probably much more likely to influence other people slightly less far out on the spectrum than them, and then those people are more likely to influence people slightly further out, and so on and so forth.

Seems obviously flawed to look at a complicated network and only look at the effect of direct influence.

Maybe bad actors within our borders are the perpetrators and the Russian bots are only repeating what a bunch of idiots in America already believe