In this week’s episode, Katie and Chris tackle news media’s usage of data. Is what you read believable? How would you go about proving it? We examine some recent claims in Bloomberg and Business Insider about racial slurs on Twitter and the process for verifying that claim, extending it to the role data-savvy organizations should play in ESG and social good. Tune in to learn more!
Key points:
There is still merit to the idea of the citizen analyst, but it is more difficult to achieve in the post-factual world. People are more likely to share information without fully reading or understanding it.
If you see data being used heavily in a news story, your first step should be to validate whether or not the news source itself is trustworthy.
If you have the capability, or if you have friends or colleagues who have the capability, it might be worth commissioning your own extraction and analysis of the data to see if the news story is accurate or even close to accurate.
[podcastsponsor]
Watch the video here:
Can’t see anything? Watch it on YouTube here.
Listen to the audio here:
- Need help with your company’s data and analytics? Let us know!
- Join our free Slack group for marketers interested in analytics!
Machine-Generated Transcript
What follows is an AI-generated transcript. The transcript may contain errors and is not a substitute for listening to the episode.
Christopher Penn 0:00
In this week’s In-Ear Insights, can you believe what you read in the news? So this past weekend, there was a news story that I saw in I think was either Bloomberg or Business Insider saying that on Twitter in particular, the increase in the usage of various racial and and other slurs was up.
I think the new stores had 500%, since the takeover of Twitter and new under under the new management.
And as soon as I read that, I said, is that true? Like? Could we verify it? Because see, yeah, I mean, sometimes you see a lot of stuff in the news being very performative sort of click Beatty stuff.
And so I asked the question, is that true? So Katie, when you see a story in the news, particularly story where there’s a big data point as the headline, what do you think?
Katie Robbert 0:58
Oh, I mean, the news like, I usually err on the side of I don’t know that that’s true until I can see other similar or authority, authoritative sources reporting the same kind of information.
And so obviously, I saw, you know, your analysis over the weekend, Chris.
But I also saw that same analysis, being picked up by larger publications from a quote unquote, third party, Social Media Research Agency, which I had never heard of before, which isn’t just saying anything, I haven’t heard of everything.
But it was one of those, like, if this is being picked up, and this is such a big deal.
Has anyone ever heard of this agency before? You know, and I didn’t dig too far into the article, because it was pretty depressing.
But you know, it’s what was the, you know, the first thing I would look for, if I was trying to validate it myself is, you know, have they been explicitly clear about the methodology? Do they have a website that helps shed some light on who they are, what their values are those kinds of things? You know, so if I’m looking at a data point from an agency I’ve never heard from before, that’s a couple of the things I start to look for first, and also, which publications are picking up this particular set of data? So for example, I first saw that analysis, not yours, the other one on Yahoo News.
Now, I don’t know if Yahoo News is a trustworthy site or not.
I mean, I think of Yahoo, as, you know, as outdated as AOL.
But I could be wrong.
So I certainly have some research to do in this in that instance.
You know, I don’t have the skill set to rerun the analysis myself.
But that’s something that I would probably assume would be a good idea for most people who have that kind of resource.
Christopher Penn 2:57
Well, and that’s a really good point.
So this is the data that we pulled.
So what we did is we queried the Twitter API directly said, Okay, here’s this, this one offensive racial slur, just show me the number of raw occurrences.
And you can obviously see, in the five days preceding the weekend, it still occurs, obviously, it’s Twitter’s a big place.
But then you can see a definite a typical pattern that emerged when the story first came out was on Friday, Friday night, and then into Saturday.
And then, of course, you see the huge spike on Saturday, and as a small but still significant spike on Sunday.
So this raises the very interesting question.
If you want to validate a news story, obviously, they have the skills and the ability to query an API directly and do data visualization, you should because it’s, it’s a good idea to get in the habit of validating news sources.
Because again, a lot of news companies make their money on page views and page views have to come from inciting strong emotions.
But what if you don’t have those skills? Like in your case, Katie, you don’t have the ability to write our code and stuff? How would you? How would you go about validating or trying to validate if that’s real?
Katie Robbert 4:18
Well, as I had mentioned, I would start to look at the source of the, you know, the data analysis, you know, is this a company that anyone’s ever heard of before? So I do have some ability to do research on companies and sort of see, like, are they legit, you know, what kind of traffic do they get? What kind of inbound links do they get? Who’s linking to their stuff? So I do have those resources through various SEO tools.
And I feel like you know, if you are, you know, a low level marketer like myself where you know, you know the basics, you can use the tools, that’s a good place to start because you can at least validate that a source of news or website is, you know, they, it’s not just something that somebody stood up that day just to sort of like, publish some false data.
And so you can start to validate it that way.
And then you can start to look at is there anyone else talking about this? So a lot of the search a lot of the social media tools have that search built in.
So start looking for that particular data source like that agency, or whoever created it and see like, are they a legit agency are other people you know, resharing, this, who are the people who are resharing, this, and so it’s interesting, Chris, you know, so if you pull up your analysis again, and if you’re just listening to this, you can see this on trust insights.ai/youtube.
That what strikes me and a lot of the conversations that I saw on Twitter over the weekend, were people quote, unquote, testing the system in terms of censorship.
And so part of me wonders, if that anomaly spike was people trying to see if they truly were allowed to say whatever they felt like, and they were just using one of the most, you know, hateful and offensive terms they could think of, in order to see if the new owner of Twitter would actually ban them or not.
Now, that’s not to say that that’s an okay way to test the system.
But based on a lot of the chatter that I saw, I’m wondering if that’s what the anomaly is.
And unfortunately, an analysis like this doesn’t give that context to see like what the words were around it.
And that would be the next piece of information I would start to look for is what additional context exists for an analysis like this.
Because if you were to put this out there, and just say, and this is what’s on Twitter, right now, with no additional context, I think that would be very irresponsible, because you’re not saying that, you know, the red line on this graph represents a chain of ownership, the conversations started to look like this, these are the, you know, near terms around these other, you know, hateful words.
And so getting a better understanding of what this actually represents, versus, you know, everybody in their mother started to use the N word, and this is what happened.
And if that’s the case, that’s fine.
But I feel like you would need more evidence to prove that that’s what was happening.
Christopher Penn 7:23
Right.
And it’s interesting, late last night, the interim head of the trust and safety team said that no, no moderation rules had changed, and that they that really big spike was a bot storm.
It was a bot network network.
However, the other spikes are not necessarily that it’d be the case, to your point.
And and this is something that, you know, we have that data, like we collect the raw data itself, and one of the things that I’ve been thinking about is, do we just take that data, bundle it up and put it out, as you know, as a GitHub repository thing? Anyone, anyone is welcome to look at the raw data.
Because analyzing hate speech is not one of our services.
As a company, you know, certainly something I did obviously did on my free time.
But it’s also not something that I want to invest, you know, hours and hours and hours developing more advanced language processing work, but I could open source it and say, here’s the data do with it, what you will have all of this is from the public timelines, this is not does not contain any private data, it does contain person identifying information.
But to your point, you’re publishing this because we published I published this thing this Saturday, Friday, I said, I don’t remember even now when I published it, but it obviously did get a fair bit of attention with with the data we had available at the time.
In those cases, where you are somebody who cares about a topic or something that you see in the news, if you don’t have those capabilities, and you’ve done the basics of what you can write, you’ve said like, Okay, we’ve we’ve looked into, you know, whoever this Chris Penn character is, and this Trust Insights company, okay, they look like they’re a real company.
What’s your next step for saying, Can I, how do I validate this is just going to the person who posted saying, Hey, can you share your sources? Or how do you go about digging deeper without a coder on hand?
Katie Robbert 9:26
Well, I think that that’s exactly it.
So if, let’s say I was the editor of a major publication, and I saw that, you know, this guy, Chris Penn, shared this really interesting data point, and I wanted to cover it.
That would be my first stop in terms of due diligence was like, Hey, let me talk to the person who created this and say, How did you do this? What are your sources? And so I actually did see a lot of people in the conversation asking about your methodology like, what does this contain? Did you know was it this was it that and I felt Like that was really responsible on their part, to get a better understanding of the context behind.
Because again, when you’re looking at this without any context, you’re just looking at basically, single data points plotted day over day.
And you don’t really have any other information, and it can tell his story, but it doesn’t tell the whole story.
So that’s where I would start.
So let’s say, you know, I saw John posting some really interesting data on a social network or somewhere else.
My responsibility would be to say, Hey, John, can you help me understand how you Where did you get this data? Did you, you know, do it yourself? Did you borrow it from someone else? You know, if I wanted to use this, how would I cite it? What would be the methodology that I would put because every time you reshare data that you don’t know where it came from, you know, you risk you run the risk of sharing false information.
And so you yourself, your personal responsibility, is to make sure that you understand, you know, can I trust? And can I validate the data that I’m looking at? How can I believe what I’m reading, and that’s it, you need to do the work, and not just blindly share based on a headline.
And I think that that’s one of the things that a lot of people have been doing over the past few days, or even, you know, since the dawn of, you know, forever, you know, just re sharing information without fully reading the article or fully understanding the whole story.
They’re just, you know, spreading gossip and, you know, just clicking the like, and reshare button without really understanding, is this something I meant to reshare, I think the same thing happened.
When Twitter did a massive round of layoffs, there was a lot of retweets and reshares, without fully reading, like the fall thread or something, I saw a couple of those posts of like the oops, I don’t have an undo button.
Christopher Penn 11:55
About 10 years ago, IBM came up with this idea of the citizen analyst and was later pivoted into sort of the citizen data scientist.
And it was a concept that was good.
It was good idea.
But it never took off.
The idea that ordinary people could pick up analytics tools, and do stuff like this, you know, when you see a news story validated, do you think do you think there’s still merit in that idea? And how realistic is it? Especially since we are now in what Jay Baer recently called in one of his newsletters, living in the post factual world where facts don’t matter anymore? Just opinions?
Katie Robbert 12:32
You know, I do think it still matters, because I think that empowering people to do their own research is a good thing.
Now, here’s where there’s that that Asterix disclaimer of, you know, just because I look at this data, and you look at this data, Chris, doesn’t mean we walk away with the same insights or takeaways.
And I think that that’s the disclaimer.
So let’s say I had the skill set to pull this data myself.
And let’s say and this is full disclosure, totally not true.
This is a made up example, for the sake of the conversation.
Let’s say I was a roaring racist.
And I looked at this and said, Okay, this is great, this is good news.
This is positive, you and I just looked at the exact same data set and had two very different reactions.
And so I might look at this and go great, now I can go get all of my, you know, similarly racist friends, get them to join Twitter, because now, this is a great place where I can be spreading this hate speech.
And again, that is a completely made up.
fact about me, that is not true at all.
I actually am really sickened by what I’m seeing.
I think that it’s a horrible thing.
But for the sake of argument, that is a made up example.
So I feel like yes, the idea of a citizen analyst is a good idea, with the caveat that it can be a positive and a negative and negative in the sense that just because somebody can get their hands on the data doesn’t mean they will do with it, the exact same that you will do with it.
Christopher Penn 14:09
Yeah, I think the two obstacles that the idea ran into one was the skills right, because even with the best tools, the skills to do data analysis, well are, are very difficult to teach.
Particularly you methodology, rigorous statistical relevance and things like that.
Those are all things that are critical to doing good data analysis.
And that requires education.
And to the point you raised is a really important one, which is in, like I said, in Jay bears post factual world.
What’s more likely to happen is that people will cherry pick the data that tells the story that they want to have told you not even in the example you’re talking about just saying whether or not this is a victory for free speech or not, which is what Um, some of the folks are saying about it.
You, that’s not even manipulating the data.
And a lot of the cases, they’re people who have em, this is true of news publications to, people will manipulate data to fit the story rather than manipulate the story to fit the data.
And that level of ethics, professional ethics is, again, something that the citizen data scientists, the citizen analysts will not have, just like the citizen blogger, or the citizen YouTuber, does not have the same constraints that a professional publication does a person can go up on YouTube and say, all kinds of crazy things like the world is flat and not round.
And there’s no consequence to this, because they do not have any of the same journalistic integrity.
However, what we’re seeing in the media landscape, obviously, is that seems to matter less to folks who want to see and hear things that agree with their point of view.
So that’s, that’s, you know, some of the reasons why the citizen data scientist stuff never really went anywhere, do you think there’s a place for, I don’t know, like a guild or a volunteer organization, or something that would fulfill the same role, you know, not asking individual randos to try and do generalized additive models, but professionals who want to do something in their spare time? And we’d be willing to say, yes, we’re going to sign off on doing this, right, we’re not going to get paid for it, but we’re going to do it right.
So that when somebody sees a news story, there’s, there’s, I wouldn’t call professional fact checking, but there’s, there’s somebody else providing some validation of data points.
Katie Robbert 16:49
Well, and I think that that’s sort of the purpose of those, you know, third party research firms, that’s what they’re meant to do.
And so, you know, it’s, you know, it’s interesting, because it’s the exact same challenge of, you know, humans programming, artificial intelligence and algorithms, unless you have a fully representative committee that is unbiased, and or willing to understand their biases, then you’re never going to get, you know, the perfect algorithm or the perfect, you know, dataset or training data.
And I feel like this is the exact same case.
And so, who then gets to validate that these people, these volunteers, for this Civilis citizen analyst committee, are then, you know, not having their own biased and, you know, secret hidden agendas when they’re publishing the data? And so it is, it’s a really tricky question that doesn’t have a great answer.
And so the, I would say, you know, and I don’t mean to, for this to sound sort of, like disheartening, but the best that you can do for yourself, in terms of whether or not you can trust what you’re reading, and, you know, validate the information is to do your own due diligence.
And that doesn’t mean you know, pulling out your machine learning, you know, programming, coding things, but it’s really just understanding who’s behind putting this data together? And Are they someone that you feel like, you know, if you were, I don’t know, gosh, you know, asked to speak about under oath, like you would feel comfortable speaking the truth of like, Yes, I fully believe that this data is represented in a way that I, you know, understand it, and I can stand behind it, you know, and again, it comes down to sort of your own personal beliefs, like, in that really terrible example, Chris, we could be looking at the exact same data set, and have very different beliefs about it.
But at the end of the day, the data itself was correct.
And our just our opinions in that bad example, just would have been different about what the data is saying.
Christopher Penn 18:56
We need to get like headbands or hair bands for ourselves that have devil horns, so we can have like a be evil version of ourselves.
Katie Robbert 19:04
Yeah, so the fully represent like, Okay, so now I’m pretending that I am the evil opposite version.
Christopher Penn 19:11
Exactly.
A goatee.
Like that’s what they do in Star Trek,
Katie Robbert 19:13
something.
But well, that wouldn’t work for me.
But I do again, sort of, and now I’m like, nervous, like full on disclosure, like that was a made example.
No, we get that.
I know, you know that, Chris, but I want I’m sure everybody knows that.
Christopher Penn 19:29
We’re getting devil here.
What is the obligation of professional organizations like Trust Insights as a company with societal topics like this? Because there are if you look at the most recent cmo survey, there was something like like 40 ish percent of organizations have CMOS that said, they don’t want their brands getting involved in anything even moderately polarizing.
As from your perspective as the CEO of the company, what is our if any our professional obligation toward stuff that ain’t gonna make us money.
But does some level of social good.
Katie Robbert 20:07
I think the the data that you shared makes sense because at our core, we do natural language processing, we do data analysis, in order to help people make their own decisions.
So I don’t feel like that was out of line in terms of what we do now, again, is it a service? Yes, natural language processing is a service data analysis as a service, and at the end of the day, data is data.
Now, as companies are trying to decide their position on, you know, these kinds of conversations, it you really need to take a step back and say, Do I have a place in this conversation? I, as a human, probably have strong opinions, you know, one way or the other about the conversation? But do my opinions align with the mission and values of the company that I work for? Or I own? Or whatever the case is? You know, and will my opinions in the conversation add any value? Or am I just in an echo chamber of everybody else is saying the same thing.
And I want to, you know, just sort of like, stick my finger in there and go, yeah, yeah, that’s me, too, which doesn’t really add any value.
And I think so it’s tough, it’s a really tough place right now for companies to navigate in terms of knowing what’s right and what’s wrong, because a lot of people don’t want to hear from the companies.
Unless you’re a company like Adidas, and you have a very public person, who was saying things like, I can say, whatever I want, I can say, anti semitic things, and I will not be let go from this contract.
And everybody was calling for Adidas to make a decision.
And they did, they finally did.
And unless we work there, we don’t really know what happened behind the scenes.
And so that was a very clear cut case where the company needed to make a decision and say something publicly.
But in this case, you know, Twitter, you know, we can all have opinions about it, but Twitter are the ones who need to make the decision.
And unfortunately, the person who’s in charge is the one who’s fostering all of this hate speech and negative conversation.
And so we as companies, we as individuals, can choose what we do with the information, we we don’t need to propel the negativity any further.
Christopher Penn 22:35
You said something interesting in there, about companies and, and sort of the brands and the values and things like that.
In a post factual world, where opinion matters more than fact that we’re that, you know, having those points of views really determines how people see the world.
Do you think that? Do you think that is what is driving consumers to make more value based choices? For brands say, like, I will not, you know, shop with this brand and stuff? Because they don’t reflect my values anymore? And because now, in a post factual world, I place a lot more value on my opinion.
And, therefore, which brands align with it and which don’t? And if so, then does that make doing this kind of work more of a strategic imperative? after all?
Katie Robbert 23:33
I think so, I think that, you know, brands don’t necessarily need to get involved in, you know, all of the noise that’s on a social platform.
However, they do need to make sure that their mission and their values are clearly stated in their assets on their website that they are, you know, following them with all of their actions.
And so when someone a consumer, you know, someone who’s looking to hire that brand, that agency starts looking at, okay, what is this a company that believes in sustainability, but also, you know, behaves in a way that they show they believe in sustainability, for example, okay, then that’s a company that I can get behind.
Versus this is a company that says that they believe in a greener world of sustainability, but yet they are still underpaying factory workers in you know, terrible conditions.
So clearly, you know, words written on their website have no meaning.
And so again, it goes back to you as the consumer needing to you to do your own due diligence and the responsibility of the companies is to not only talk the talk, but walk the walk and your actions as a company, need to reflect the things that you say you’re publicly gonna do.
So for example, if you say in your values, we will, you know, publicly come out against hate speech, then yeah, you need to be part of that conversation.
But if your values are, we believe in a more sustainable world, maybe you don’t necessarily need to partake, you just need to keep sort of like toeing the line and moving in the direction that is appropriate for you.
Christopher Penn 25:15
So for us, obviously, me as an individual person, they’ll have some level of curiosity in this particular set of data.
What, what, if anything, would you say would be sort of next steps? You know, is this just a curiosity that blows over with the news cycle? And, you know, whatever happens today, and news takes precedence? Or is this something that becomes part of part of what I do individually, or we do as a company?
Katie Robbert 25:44
You know, I think that based on what we do as a company and the things that fall into our values, it’s not inappropriate for us to at least continue to educate.
On here’s what the data looks like, here’s a summation of the conversations, I don’t think that that’s an appropriate at all, as long as we don’t, you know, take it too far down the line and start telling people how they should feel about it.
Our job, our goal, our company is built on educating people on what the data says not telling them how to feel about it.
And so I don’t think that it’s an appropriate for us to continue to do that.
I do think we don’t we want to make sure we don’t cross that line and say, and this is exactly how you should feel about it as well.
Like, that’s not our place at all.
Christopher Penn 26:33
Right? Yeah.
Which which I reflected, I think in the original post, which was, here’s the data.
I’m not gonna tell you how to do it, how you feel about it, but here it is.
Katie Robbert 26:41
Right.
And that’s exactly it, that’s our responsibility is helping people understand what’s happening.
Christopher Penn 26:48
All right.
So to wrap up, if you see data, being used heavily in some kind of news story, your your first step should be to validate whether or not the the news source itself is trustworthy.
Whether the data provider is trustworthy, you’ve ever heard of it.
And then if you have the capability, or if you have friends or colleagues who have the capability, and you’ve got the time, and it’s something you really care about, it might be worth, you know, commissioning your own extraction and analysis of the data to see if a what’s in the news story is accurate or even close to accurate, and then be what are the nuances that didn’t get reported.
And if this, if this topic, in particular, of hate speech, within social networks is something of interest to you, and you want to talk more about it, pop on over to our free slack group where we talk about analytics of all kinds and sidled stuff social good stuff is not off the table as long as it’s done in a respectful manner.
Hop on over to trust insights.ai/analytics for marketers, where you have over 2800 other marketers are asking and answering each other’s questions every single day.
And wherever it is that you choose to watch or listen to this show.
If there’s a challenge you’d rather have it on instead, go to trust insights.ai/t AI podcast if you liked the show, please leave us a rating or review.
We appreciate it.
It helps helps more people find the show.
Thanks for tuning in today and we’ll talk to you soon
|
Need help with your marketing AI and analytics? |
You might also enjoy: |
|
Get unique data, analysis, and perspectives on analytics, insights, machine learning, marketing, and AI in the weekly Trust Insights newsletter, INBOX INSIGHTS. Subscribe now for free; new issues every Wednesday! |
Want to learn more about data, analytics, and insights? Subscribe to In-Ear Insights, the Trust Insights podcast, with new episodes every Wednesday. |
Newsletter
Podcast
