A new study is out that tries to measure the effectiveness of social media advertising campaigns in political races. The unique part of this study is that it makes use of an actual advertising campaign during the 2020 presidential contest that deliberately held out a control group so that its effectiveness could be measured:
We present the results of a large, US$8.9 million campaign-wide field experiment, conducted among 2 million moderate- and low-information persuadable voters in five battleground states during the 2020 US presidential election. Treatment group participants were exposed to an 8-month-long advertising programme delivered via social media, designed to persuade people to vote against Donald Trump and for Joe Biden.
The funny thing is that I think the authors underrate their own results. For example, here is turnout for Republicans and Democrats:
The authors say, "We found both small mobilizing effects among Biden leaners and small demobilizing effects among Trump leaners." But this is a net difference of 1.8% in turnout. In most political campaigns this would be considered pretty substantial and the price tag of $8.9 million for five states pretty modest. Most campaign managers in battleground states would be thrilled with it.
Basically, I think you can say two things here. First, on an absolute basis this study shows a fairly small effect. Second, within the context of a close political race, it shows a very substantial effect.
Exactly the argument of those Democrats who attempted to pass voter rights legislation in the last Congress. 1.7 % is huge in a close election, and partisan discrimination could easily result in similar percentages.
How many 2022 elections were decided by 1% or less? With such legislation might we have had a Senator Mandela Barnes in Wisconsin?
But filibuster rules were deemed more important by two Democrat Senators and every single Republican.
Start making more money weekly. This is valuable part time work for everyone. The best part ,work from the comfort of your house and get paid from $10k-$20k each week . Start today and have your first cash at the end of this week.
Visit this article for more details.. http://incomebyus.blogspot.com/
How are "moderate- and low-information persuadable voters" defined? Are there any demographic biases? From the error bars, it looks like most of these voters were "other" in partisanship and they didn't move at all. Assuming half these people are "other", then the change looks to be less than 1%. Am I reading this correctly? Back to my original point, what proportion of voters are moderate- and low-information persuadable voters?
The effects are at the margin--therein lies the rub.
The error bars overlap. Not by much, so odds are the difference is real, but you can not definitively claim the differences are real. The limited claims by the authors are appropriate.
Here's the thing, I don't know if these type of studies could get much better than this.
For thought exercises, substitute in different studies for the one at hand and see how the perception of these results affect your opinion.
1. Results are for new Alzheimer's drug.
2. Results are for side effects of a new vaccine.
3. Results are for a link between cancer and air pollution.
4. Results are for the health benefits of drinking coffee.
Note: major breakthroughs can drive businesses, health, etc., but many times it's the incremental improvements that end up winning the day.
Not sure it is doable, but you need to measure the effect across the entire electorate, not just the voters who are shown the campaign. For example, look at something like "her emails," that went viral. I bet for each person who read some dopey Tweet about that the story was repeated ten times at the barber shop, or in the gym, or on the bus. I share Tweets all the time with non-Tweeting friends. Facebook nonsense is even worse.
There are too many variables to draw any valid conclusions from such a project. It wasn't "social media" that was effective, but the complex relationship between the material posted on social media, the people being targeted, and the broader social/political/economic/media context. Nothing in that very complex relationship will ever be replicated exactly in another campaign; on the contrary, most elements will be quite different.
Not just elections.
Hate speech rises on Twitter in its largest markets after Musk takeover
With 75 percent of its audience outside the U.S. and Canada, the impact of Elon Musk’s moderation cuts has been great elsewhere
https://www.washingtonpost.com/technology/2023/01/14/twitter-moderation-cutbacks-impact/
"Musk cut virtually all staff in Brazil, allowing an unmoderated surge in misinformation that helped fuel this month’s attacks on the country’s government center."
Come to the ‘war cry party’: How social media helped drive mayhem in Brazil
Researchers detected a surge in aggressive rhetoric from election denialists in far-right channels online ahead of Sunday’s rioting
https://www.washingtonpost.com/technology/2023/01/08/brazil-bolsanaro-twitter-facebook/
"In the weeks leading up to Sunday’s violent attacks on Brazil’s Congress and other government buildings, the country’s social media channels surged with calls to attack gas stations, refineries and other infrastructure, as well as for people to come to a “war cry party” in the capital, according to Brazilian social media researchers."
Good information. Thank you.
Many of the comments rightly point out that the results of this study are on the border of statistical significance, and therefore much more rigorous study would be needed for proof.
Well, there either is a significant effect, or there is not. We now have some evidence, inconclusive as it is, that there is a significant effect. If we wait until there is unambiguous evidence before taking action, we would likely wait forever. Proof is a very high bar.
In general, I feel as if there is some valid evidence that points to a significant conclusion, it is often wise to act on it rather than wait. "Maybe" is not proof, but in life "maybe" is often the most you will ever get. It moves the probabilities in your direction, and "maybe" might lead to a good outcome that would be absent had you not acted on the basis of some evidence.
I have had some experience of this in a medical context. Doctors will not act on a treatment except on the basis of proof. Patients are more interested in possibility and likelihood on the basis of limited evidence.
I would say the only valid conclusion that can be drawn from the exercise is the tautology that social media campaigns will influence some voters if done in a way that's effective in influencing them. The fact that this particular campaign appears to have had a modest influencing impact is no help at all in designing a campaign that will have an impact in a different context.
I think the problem is we have become so lazy that we no longer WANT to fact check
How many people form their opinions from something they read on social media? We have developed this societal confirmational bias that has affected so many of us.
We USED to be able to trust the news on the major networks because they had journalists doing the reporting. Now we have limited journalism but LOADs of opinion schleppers that we, as a society WANT to believe are news people. Hannity, O'Reilly, Ingraham, Carlson, and the group at MSNBC.
And of course the internet is the main culprit. How many of us would be writing letters to the editor at Mother Jones if it was print only and cost $69 a year to complain or compliment Kevin Drum?
It is far too social and less and less like the 'classic' media we all grew up with. Gone are the days of Jennings, Reynolds and Cronkite. Add to this our lack of honesty about our own short comings and we have a recipe for a whole society FULL of ill-informed people who may vote.
So, yeah "social media" is doing MORE than influencing elections. It's influencing what we want to believe is right versus wrong.
The 3 most liberating words of the English language are "I was wrong". Nobody says this anymore. Since when did we become so brilliant as a society?
We can't tell the difference between right and wrong and news and views.......