How social networks can be used to bias votes | Hacker News Reader

How social networks can be used to bias votes

nature.com

If only this level of scrutiny was applied to traditional mass media over the last century. All of these techniques are just updated versions of handing the anchorman a biased script.

My thoughts exactly. You know what else can bias voters? Expert testimony, even in fields outside their expertise. Religious leaders. Teachers. Flyers. Advertisements. In fact, just about any form of communication by anybody. Without comparing the size of the bias to other forms of influence, this is a nearly contentless statement.

Know how you get a human to believe something? Make them work for the information. The more someone works for information the more likely they are to believe it is valid. I think this is why you see modern educated people with such deference to the "experts". They know the hard work and the costs involved in becoming a academic and as such they value that experience with an almost religious fervor.

Tying the information to their personal identity and/or local social structure seems far more effective.

The beliefs people seem to hold to most strongly, in the face of all evidence, are those that form their identity. The Dust Bowl farmer, who is a farmer, owns a farm, has a mortgage on the farm, has a family on the farm, and when the water dries up and the soil blows away doesn't know how to do anything else but farm all that much harder.

As an example -- the reference is one of the statements from a farmer faced with just that situation, and may explain a number of phenomena, including a pathological case of sunk-cost determinism.

What's your evidence that working for information helps with convincing? Because there's at least some evidence (above) that the exact opposite is true. Belief transcends all evidence and experience, or cost, including farm, family, and life.

>What's your evidence that working for information helps with convincing?

I can't say I had any hard evidence while I wrote that. I admit it's more of an observation from my experience. That said there this study[0] seems to suggest a mechanism whereby this effect could occur. Other ideas like the suck cost fallacy and effort heuristic seem to confirm that this effect has been noticed by others on some level.

[0]https://psycnet.apa.org/record/2012-08696-001

Thanks, though I don't think that's clear support (interesting research, all the same).

A study of concept/belief fixity and dynamics associated with those would be interesting. If difficult to design.

Wow. Are you really equating expertise with religious fervor?

The anti-expertise movement is what I would describe as religious fervor, not the other way round.

Real expertise isn't a problem. A expert plumber is shown to be so by his actions. Nobody would install a leaky faucet and be able to call himself an expert.

An expert psychologist for example? Her patients can remain depressed, mutilate themselves, kill themselves and even kill others and yet she will remain an "expert" mostly on the basis that she was educated at a place that similarly ineffective experts have claimed is a good place for a psychologist to come from.

I suppose the trouble is experts by education got lumped in by experts by results at some point down the line.

here, have some economics experts:

https://www.americanprogress.org/issues/economy/view/

https://www.cato.org/research/growth-development

how big is the middle part of the venn diagram here?

being skeptical of domain-experts is only 'religious fervor' if the field is cut-and-dried. many aren't

Let me guess. You cherry-picked some cases where experts were wrong. Am I right? Here, read this: https://www.amazon.com/Death-Expertise-Campaign-Established-...

I don't have time for this.

cherry picked? yes and no.

These are both hugely influential policy think tanks which hire experts. They both have hugely differing views on the economy.

Or is providing evidence to support a claim now 'cherry picking' to be dismissed? And if so, aren't you proving my point?

I don't think it's contentless, because for a long time there was an idea that Social Media would free the general public from manipulation by a small group of wealthy/powerful individuals/corporations. This article is effectively implying that Social Media is no different to the influences you cite.

Why would it be different? It is about how the brain works, and it applies across several forms of media.

I personally held the (probably mistaken) belief that by democratizing publishing information, it would be harder to create a false signal. Ie, if everyone has a printing press, then nobody can control the entire narrative, the way a handful of newspaper barons could.

Instead, it seems to me now that giving everyone a printing press made so much noise that we still rely on centralized authorities to sort, rate and surface those that we want to read. And it made it cheaper, in terms of both reputation cost and money, for someone to create a bunch of false signal.

The parallel to the printing was misleading, and the true outcome is much more complicated and mixed. The scale really changes how these things work, in terms of the incentives created and the sheer amount of content possible to consume.

Another possibility is that the "new" media threatens the "old" media so much that the old media feels compelled to exaggerate the problems with new media. Sure it's different. There is more noise than consensus. The next generation of humanity might well consider that a benefit. Don't expect old media to figure that out; all of their incentives pull in the other direction.

Don't forget movie stars, pro athletes, etc.

There seems to be a concentration of political activists among these. I very seriously doubt it's by coincidence.

What, I wonder, is the difference between social networks, and old school political machines running on non-digital social networks, or mass media networks and channels, or even conventional groundwork campaign techniques.

Does a organizing with members of key organizations count as biasing a community. How about a campaign sign placed at a strategic intersection? Or oft used last minute mudslinging postcards. For that matter what is the difference between biasing a voter, and convincing a voter to vote (maybe in the papers case on team-dynamics instead of issues, but voters often decide their votes not on direct issues...).

I fully believe though that there are techniques, accessible by money, which yield more efficient results of voter persuasions than other techniques - but those tools are not limited to social media.

If you haven't seen this -

https://www.youtube.com/watch?v=dguiAWrUGMM

And it's still going on today. Not just some bad practice they did years ago.

https://www.youtube.com/watch?v=487CRdRHEeI

Think about the control you can nudge things in a specific direction with that.

I had begun to initially post a rebuttal to OP. My rebuttal was going to be. This is different since today one could

1. Figure out people’s political leanings via their data

2. Figure out target networks to influence

3. Plant influencers

Have watched the videos. I had not realised the media in the USA was that localised and co-ordinated. Initially thought it was some kind of prank/joke between the stations. With this level of co-ordination and functioning a specific (untrue) narrative could be spun. One could even a/b test and see what has the most desired effects.

Chilling. The question then becomes: What narrative is being spun? Who does it benefit? Who does it disadvantage?

What do you mean "could"?

https://theintercept.com/2019/01/20/beyond-buzzfeed-the-10-w...

The problem is not only do people “want to believe” but actual journalists who went to journalism school get trapped into the “wanna believe” trap coupled with not wanting to miss out on a scoop results in terrible “reportage” and results in a kind of weird propaganda.

clearly this is russian propaganda. clearly.

Manufacturing Consent: The Political Economy of the Mass Media (1988) by Edward S. Herman and Noam Chomsky does talk about how the state uses mass media to influence collective opinions.

From https://en.wikipedia.org/wiki/Manufacturing_Consent

> ... the authors propose that the mass communication media of the U.S. "are effective and powerful ideological institutions that carry out a system-supportive propaganda function, by reliance on market forces, internalized assumptions, and self-censorship, and without overt coercion", by means of the propaganda model of communication.

And, you know, religion..

> We initially assume the payoff to the losing team is positive, b > 0, which reflects a ‘compromise worldview’: it is preferable to reach some decision than to end in deadlock. This payoff scheme captures the common practical value of broad consensus in collective decisions, even as individuals pursue partisan preferences12,13. There is ample evidence that large majorities of Americans, for example, adopt a compromise worldview in their attitudes towards political decisions14. However, others adopt a ‘zero-sum worldview’ in which they prefer deadlock to their party losing, that is, b < 0. The behaviour of people with a zero-sum worldview is simple: they act as zealots, meaning that they always vote for their preferred party regardless of the poll that they see. We begin by focusing on players with a compromise worldview, and then we study groups with a mixture of compromise and zero-sum players, including zero-sum bots.

In the context of voters in our modern politics, does the "compromise" viewpoint even exist?

They do, but they're decreasing in number. They're also kind of boring so they don't get coverage the way highly vocal 'zealots' do.

Data on political polarization in America is really interesting. One feature of the data I find fascinating is that in 1994 many people were not "consistently liberal" because they opposed gay marriage. With the big shift toward acceptance of gay marriage these people are now considered "consistently liberal" and by Pew's reckoning this means that liberals have become "more polarized". In my opinion there is no morally defensible compromise position on whether homosexuals deserve their natural rights, so I find this quirk of the survey interesting.

If you look at the ten questions of the Pew survey, it seems to me that anyone who can answer on the conservative side of all ten questions (in order to earn the "consistently conservative" epithet) must be the biggest jerk on Earth, whereas someone who answers on the liberal side of all ten has sensible and morally justifiable views, hence I take all the hand-wringing over polarization with a grain of salt.

https://www.people-press.org/2014/06/12/political-polarizati...

I don't want to nitpick, but I don't see how gay marriage in this context is a natural right, since really it is just a government actively providing a service(marriage licensing). A natural right might be more like the right for two gay people to cohabit, or a religious marriage joining two gay people in marriage.

The actual question on the Pew survey is "Homosexuality should be discouraged by society." agree/disagree. The fraction of people agreeing with this statement has fallen greatly in the last 25 years, mostly due to demographic replacement.

You can discourage something(which is a natural right of free speech), but still think people have the right to practice it - you might just think it is a bad idea for individual or social health(e.g. You can discourage people from becoming hermits, because a society of hermits isn't much of a society at all)

It's an interesting study in terms of how they conducted it, but isn't it obvious? If political ads didn't swing votes, no one would run them.

It's not at all obvious that political ads (or ads in general) work. The correlation of political ads (and all other kinds of political stumping) with success is confounded by donations (because more popular candidates are able to get more donations), and looking at randomized political interventions, we see tiny effects - you might've seen some coverage of "The Minimal Persuasive Effects of Campaign Contact in General Elections: Evidence from 49 Field Experiments", Kalla & Broockman 2018 https://web.stanford.edu/~dbroock/published%20paper%20PDFs/k... which says bluntly: "a systematic meta-analysis of 40 field experiments estimates an average effect of zero in general elections." Political scientists often find minimal effects of advertising spend in general (how much did Trump spend compared to Hillary? to use a particularly pointed example). And to the extent that some kinds of ads do work, they may be self-negating ('advertising wearout' and political examples in https://www.nytimes.com/2018/05/25/business/what-advertising... ). Ultimately, 'if people don't want to go to the ball game, how are you gonna stop'em?'

More broadly, there are many experiments on online advertising showing that the correlational biases are huge and you get wildly misleading estimates if you use anything but rigorous randomized experiments, and the actual causal effects appear to usually be, again, close to if not indistinguishable from zero, which I link in https://www.gwern.net/Ads#discussion It is extremely easy to fool yourself in this subject.

For an even broader argument that we greatly overrate the manipulability of people, see Mercier 2017's "How Gullible are We? A Review of the Evidence from Psychology and Social Science" https://www.gwern.net/docs/psychology/2017-mercier.pdf .

I don't think anyone would suggest that there's always a linear correlation between ad spend and conversions.

But are you saying that ads never influence decisions or actions? That seems.... impossible. Ads must work sometimes. The degree to which they may or may not, however, is extremely hard to figure out, as you note. The study that this HN thread is about doesn't shed much light either, other than saying that "people can sometimes have their decisions influenced." Same as always.

People can certainly change over time. 30 years ago, gay marriage would've been inconceivable by the majority of the US population. How did that change? Many of those same people who were against gay marriage are still alive! They must've been influenced by something - advertisements or not.

> But are you saying that ads never influence decisions or actions? That seems.... impossible. Ads must work sometimes. The degree to which they may or may not, however, is extremely hard to figure out, as you note. The study that this HN thread is about doesn't shed much light either, other than saying that "people can sometimes have their decisions influenced." Same as always.

Nobody is saying that ads have literally 0.00000% effect. (I would strongly disagree with such a claim: https://www.gwern.net/Everything )

What I am disagreeing with is the claim that they are so powerful and so obviously pass a cost-benefit test that they are known to work and further, that the mere fact that people buy ads is sufficient to prove that they are highly effective and worth buying. ("Isn't it obvious? If political ads didn't swing votes, no one would run them.") All of these claims are quite likely to be wrong: the best evidence is that the average effect is minute (and one can't rule out even negative effects the opposite of desired), they probably do not pass a cost-benefit test because much commercial advertising does not, it is easy to be mistaken about their value, and given how 'obvious' it is that they work and which casual analysis would support such interpretations ('candidates who spend more on ads win more! isn't it obvious that ads work?!') we should not be surprised if people buy advertising even when the true effects are unknown, unknowable within reasonable possible experiments, and probably minute.

That's totally fair. I was being a bit flippant in my original comment.

I was reacting to the idea that, a candidate spreading their message will impact what decision people make at the voting booth, and that in and of itself is obvious. If candidates said nothing, no one would vote for them, because no one would know anything about them. Surely that much is obvious.

Yes, it's probably very difficult to prove the cost-benefit analysis of ads. At best, people hope that attack ads will reveal some information to people that will sway their decision (I perused through the papers you linked, but didn't see much about scandals). I also imagine ads are meant to generate buzz in the media, which can push a message to would-be voters, who can then keep spreading the message. The effect is likely over time, and extremely difficult to predict & prove due to many different ways people can interact with each other.

Although that paper you linked on gullibility was..... odd. I don't know why the distinction of "content-based" gullibility vs. "source-based" gullibility is important or interesting. The authors are saying that.... fake news works? That also implies that advertising can work if it's delivered in a medium appearing as an authority, i.e. a media organization or research group.

In other words, bad advertising doesn't work, but good advertising will. Surprise?

> 30 years ago, gay marriage would've been inconceivable by the majority of the US population.

Gay marriage wasn't a major political issue 30 years ago, so I'm bet most people hadn't contemplated it.

> Many of those same people who were against gay marriage are still alive! They must've been influenced by something - advertisements or not.

Only if you think a significant number of them that are still alive changed their minds. I'd wager most didn't, but that they've been slowly dying out.

I don't disagree that advertising has an effect, but I'm not sure that this a great example of that, because advertising tends to accelerate people down a path they were already headed in, whether to buy a product, or believe a political conspiracy, but rarely does it push people to completely reverse their direction completely.

This study would also indicate that political polling has the ability to influence elections?

I thought everyone knew that already given that push polls are a thing.

Of course, anything that brings focus potentially changes the final metric.

I would agree, but I think most political reporters and pollsters view themselves as impartial observers rather than active participants...

What they view themselves as does not change the fact that they skew things in one way or another. The only observer that is neutral is the one that is not seen and not heard of.

>All data necessary to reproduce the results are available at https://github.com/jplotkin/InformationGerrymandering

>For example, when just a few yellow-party zealots were deployed strategically among a larger number of undecided players in the purple party, these bots were able to sway the majority opinion towards the yellow party. This was true even when the parties had exactly the same number of members, and when each player had the same amount of influence.

I'd like to know a little more about the experiment, specifically how much information was available to individual voters. Did they know the votes of specific other voters? Of their party specifically? Or just the overall polling situation? Were they allowed to communicate, like through a chat box or something?

Without knowing these details I'm finding it a little hard to understand how these few 'planted' zealots were able to influence the result that strongly.

There is an ongoing controversy in Norway right now, where a TV show during one month tried to manipulate the results of "school elections" (basically mocks of the real elections, where one votes for the real parties - the results are often used as an indicator for the real election results some weeks later).

The goal of the show was to show how election results can be manipulated, which created some outrage: some of the students are old enough to vote in the real elections, so they might have manipulated real results.

Norwegian news source: https://www.nrk.no/norge/folkeopplysningen_-slik-forsokte-de...

Yes, it all comes down to money. Money = the means to influence = power. It is interesting to note that citizens have been self-inflicting political bubbles upon themselves this for hundreds of years by surrounding themselves with similar-minded people (i.e., people who aren't going to argue with their politics).

Votes are biased by worldview and worldview is tied to group identification.

i think this might cross over to something like: Election campaigns can be used to bias votes.

That being said social media has a great potential to be used for campaign violations.

Elections considered harmful.

It's impossible to tell from TFA or the abstract what precisely is meant by "information gerrymandering". One imagines some sort of graph theory or "six degrees of Bacon" argument. "This node should only influence some number of other nodes, but in fact if you count second- and third-order connections it influences some higher number of other nodes!" Presumably they've picked out some particular criteria for differentiating justifiable influence from unjustifiable influence. The criteria are probably fine; since we don't know them the point is moot.

Even if one accepts all of that, one still can't follow TFA to its conclusion of "...voters really can be manipulated in the digital age. Legislators and regulators must take note." We've been working on this "giant graph of nodes influencing each other" problem for decades now. That's just SEO. That's the basic thing that Google does. Any network has these properties, even the "snail" networks of pamphlets and broadsheets that kept our founding fathers informed.

It was completely appropriate for late-18C pamphleteers to publish and distribute according to the dictates of their own best judgment. It's completely appropriate for today's social network operators to do the same. It's completely appropriate for users to do just exactly as they please, whatever the TOS say. What is not appropriate is for "legislators and regulators" to imagine they have a damned thing to say about this process.

1A is not ambiguous on this question. The government don't get to forbid the public from hearing Trump describing to Billy Bush his techniques of seduction. They don't get to forbid the public from hearing Clinton tell bankers they should be able to regulate themselves. [0] It's a conversation because they don't get to decide everything that gets said. For three years now, a segment of the "intelligentsia" have wallowed in a mode of magical thinking about networks and bots that has left them completely uninterested in real things. If this research can help them break out of this rut, one welcomes it. However, that seems unlikely...

[0] https://www.thedailybeast.com/leaked-podesta-emails-show-ber...

Advertising works. News at 11...

Here, all the polls are skewes towards whoever the polsters support. And have been for many years..

At least they can’t blame Trumps victory on this as the polls showed Hillary in the lead until after the election.

Why Social Media (Facebook, Twitter, etc.) isn't willing to work with local media and communities? Maybe off topic but genuine question.

We have traditional media that failed to transition to Web. Same stuff in different package, instead of collaborative approach.

And we have social media that fails to deliver the promise of communities. Fails in the transition from virtual to real world.

From my experience, success in Facebook is closed groups. They are the ones with most activity and the ones that achieves real social events.

Both players (traditional media and social media) have what the other one lacks. I'm really scratching my head on why they aren't working together.

Work with how? What would that entails? What is stopping them from getting their own right now?

Contributing with collaborative tools, billing, linking communities, managing users, comments, events, etc.

Local media adds content, moderation and insite presence.

A lot of communities can benefit from this.

Events like cleaning trash, planting trees, sport events, improving schools.

I think Facebook's biggest strength is local communities. It seems to be the best place to generate interest for events and new businesses and it seems effective.

I can't imagine local newspapers are anywhere near as effective.

Local newspapers are in-site and can contribute with the roles of leader, moderator and organizer, as well as adding genuine content. All of these Facebook Users usually don't achieve in an articulated way

Reply on news.ycombinator.com