#259: Dateline Data

There’s data, data everywhere, including in the media! Data often gets collected, analyzed, published in a study, covered by a journalist, and then distilled down to a headline. The opportunities for lost-in-translation (or lost-in-simplification? Lost-in-summarization?) misfires are many. We tried an experiment—each of the available co-hosts brought some headlines that made them raise an eyebrow, and we tested our own data literacy (data skepticism) with a real-time review. The parallels to the day-to-day work of an analyst were many!

Links to the stories discussed (in order of appearance):

Photo by Roman Kraft on Unsplash

Episode Transcript

[music]

0:00:05.8 Announcer: Welcome to the Analytics Power Hour. Analytics topics covered conversationally and sometimes with explicit language.

0:00:13.9 Michael Helbling: Hey everybody, welcome. It’s the Analytics Power Hour. And this is episode 259. How many times have we seen news articles and that make data claims? And once you actually read the article, well, it seems either just like different or overblown or maybe didn’t really come to the same conclusion. And as analysts, we have to come up with representations for our data that summarise key points and takeaways in an easy to understand way. And so connecting those two things, we have to, a lot of times, give people a the headlines, and are we doing a good job of it? So we thought we’d do a little experiment. We could sharpen our skills by taking a look at some actual headlines, do a little commentary and analysis, and see what we come up with. So in our best impression of Nicholas Fehn, if you know, you know, here we go. Do you actually know who that is, Tim?

0:01:09.2 Tim Wilson: No.

0:01:09.2 MH: Okay.

0:01:09.3 TW: Of course not.

0:01:09.5 MH: He was a Fred Armisen…

0:01:10.7 Val Kroll: It was a polite laugh.

0:01:12.9 MH: Yeah, okay. It was a polite. I thought it was a knowing laugh. So I was excited.

0:01:18.4 TW: Oh, no, no. Nope.

0:01:19.0 MH: So I’ll explain it. Nicholas Fehn was a Fred Armisen character on SNL who would come on weekend update and read news article headlines, but he never actually read them. And I love that sketch. Anyways. But let me introduce my co-host today. Val Kroll, how is the windy city?

0:01:38.7 VK: It’s actually really windy today, which is not even why it’s called the windy city, but it is quite windy here. So happy to be here.

0:01:48.4 MH: It does get windy. Maybe we’ll see an article or two about that. And Julie Hoyer, how’s Cleveland?

0:01:56.3 Julie Hoyer: Oh, Cleveland’s Cleveland.

0:01:57.7 MH: I was going to say how’s the mistake by the lake, but then I was like, they’re past that now.

0:02:02.5 JH: You know, they won a recent game.

0:02:04.3 MH: They did.

0:02:05.2 JH: So, I guess we can’t call them that this week.

0:02:07.5 MH: Yes, that’s right. We’ll say no more. And Tim Wilson, how is Columbus?

0:02:15.6 TW: It’s also kind of windy today.

0:02:17.4 MH: I don’t have anything snappy for Columbus. It’s a great town though.

0:02:20.3 TW: Okay.

0:02:20.8 MH: And I’m Michael Helbling holding down the south here in Atlanta. Everybody, Howdy Doody? All right. Let’s jump into it. I was about to say everybody go vote, and I was like, this is coming out. It’s the end of the month, dummy. So, I was…

0:02:39.4 JH: Yeah. Tim would have strangled you. Howdy Doody. So that’s where you landed.

0:02:44.4 TW: Yeah. Everybody…

0:02:45.8 MH: Is that your verbal thing?

0:02:48.4 JH: It was the build of the silent. Now, I have to say something.

[laughter]

0:02:55.6 MH: That’s all the wrong. All right.

0:03:00.4 VK: Okay. Okay. I actually, cry.

0:03:03.6 MH: Headlines. Headlines, headlines. Just give me the headline. That’s what I say to everybody who’s giving me data. Just give me the headlines ’cause I’m ill-tempered and short with everyone. Okay. How do we wanna kick this off? We can just maybe go around. Let’s share one and we can talk about it. I think it’ll be kind of fun.

0:03:26.7 VK: Sounds good.

0:03:27.4 MH: Who wants to start?

0:03:28.9 VK: So mine, actually, I feel like this was one of the early ones that I saw that kind of helped inspire this episode. So I’m excited to get all of your reactions to this. Okay, the headline is, the most stream TV shows are almost entirely reruns.

0:03:48.3 TW: Interesting.

0:03:50.5 MH: Yeah. I mean… It feels like most, there’s a lot more of rerun content than there is new content. So the population that you’re drawing from is such that that would likely be true.

0:04:08.3 VK: Yeah. I mean, my first thought was literally, duh.

0:04:13.0 MH: Well, but I mean, not necessarily ’cause depending on the platform and things like that, like, streaming platforms are always coming out with new content as well. So they’re trying to make it popular. But that’s interesting on a certain level that it’s, mostly reruns that people watch.

0:04:32.7 VK: Was there a specific platform, Julie, that it was referring to, or just like in general?

0:04:39.1 JH: So it was interesting. So this actually came through from a newsletter that I subscribe to, Tim does too. But it was on Instagram. So they post like these headlines. So it comes through with this picture, and I’m not gonna give away the TV show ’cause I’m going to make you guys guess what it is. And that’s the headline that it says. And then I swipe through to the next…

0:04:58.6 MH: Is it The Office?

0:04:58.8 JH: No. And I swipe through to the next picture, and it didn’t give any other detail of like specific streaming platform Val to your point. It just had a Bloomberg’s like ranked by Nielsen’s weekly top 10. That’s the only other detail it had is like the title of the data and even the actual comments down below, it didn’t give much more of like streaming services they got this from, where the data came from.

0:05:28.1 VK: If it was all like the me TV that’s on in your dentist office.

0:05:31.2 JH: Right. Right. Does that count? Yeah. Yeah. So it said minutes spent streaming shows while ranked in Nielsen’s weekly top 10. And then the bar chart is by billions of minutes.

0:05:44.4 VK: Oh.

0:05:45.3 MH: Oh, my.

0:05:46.7 VK: That’s quite a scale. That’s quite an Nexus.

0:05:48.5 JH: Yes.

0:05:48.7 TW: Well, but there is kind of a, I mean, like Netflix has made news periodically as to what data they release and what data they don’t. And Nielsen has kind of, it’s practically in the news as to what they can and they can’t track. So there is a part of that where that sounds like it’s being stated is like, here’s an absolute fact without sourcing, but it’s really, there are a bunch of assumptions based on what data they have under the hood. So it’s like it’s a punchy headline that doesn’t necessarily answer the, you know, what the services are, what, you know, if you’re excluding a service that skews very far away from, original content or, you know, then that would be. But there was… Suits had it’s had a moment where it was kind of up there. Friends has kind of historically been up there. So.

0:06:44.8 VK: Oh, yeah. Any other guesses?

0:06:46.4 JH: Yeah, it’s got to be like, yeah, some nostalgia itch, like some of the ones that Tim was just talking about. I feel like, I don’t know. The Office was a good guess, Michael. I can’t think of something better than that.

0:06:56.3 MH: Yeah. And?

0:07:00.3 TW: No, pregnant pauses are great on a podcast.

0:07:04.1 VK: Editing. Editing.

0:07:05.4 JH: We’ll fix that in post, Tim.

0:07:07.8 VK: All right. It is?

0:07:14.4 JH: NCIS.

0:07:14.5 TW: There you go.

0:07:14.7 MH: What?

0:07:14.9 JH: Ooh. I mean, there’s like 10 different NCISs.

0:07:20.5 TW: Yeah.

0:07:21.3 VK: See?

0:07:22.1 TW: People watch that show? But now I definitely question because that’s, if they were ranking it, that is what they time frame in a subset. ‘Cause I absolutely guarantee that Friends and Suits and The Office, depending on how you slice it. So that wasn’t in the headline, though. They were just… But the inventory of content when you’ve got, yeah, so many NCISs that have run for a decade.

0:07:48.4 MH: Question for you, Julie. Did they talk about the comparison between the reruns versus new content in terms of streaming minutes?

0:08:00.4 JH: Yeah, so let me, I’ll break it down what I was able to find. It was actually kind of hard to track this down. It ended up, they just said per Bloomberg at the end of it. And so luckily I was able to find the real article on Bloomberg. It was kind of buried halfway down as like part of a larger topic. And what I found was that they said they were examining three years of Nielsen data tracking the most watched shows every week and different streaming services share of total TV viewing every month. So they were saying that this is one of the first times they looked at like streaming service data and like, I guess, non-streaming service data. But they were looking at Nielsen top 10 every month for the… It says examining three years is what I’m taking. But to your point, they never said that they were kind of like controlling for this idea of obviously reruns have been out longer, but the top 10.

0:08:49.9 TW: What the fuck is a rerun in a streaming? I mean, I watched…

0:08:52.9 JH: Right.

0:08:52.9 TW: I watched the…

0:08:53.1 JH: Does it mean I watched it for the first time, or that it’s, like, more than a year old? Like, I don’t know how they’re defining reruns.

0:09:01.9 MH: Maybe it’s syndicated.

0:09:02.4 VK: Yeah, syndicated. That’s what I was thinking.

0:09:04.3 MH: So, in other words, production has ceased on the show itself.

0:09:06.9 TW: Well, that’s not true, ’cause NCIS, aren’t they still making that?

0:09:09.0 VK: Oh, NCIS, they’re still going.

0:09:12.4 JH: I’m sure they’re up to, like, 20 versions now.

0:09:12.4 VK: So they’re a rerun, so they didn’t say anything.

0:09:14.9 MH: NCIS, Poughkeepsie.

[overlapping conversation]

0:09:16.8 MH: Does Max’s original content wind up going into syndication? Like, when does Game of Thrones become a rerun?

0:09:26.8 JH: Yeah. I’m not sure.

0:09:27.5 MH: So that’s actually… A rerun is literally a…

0:09:31.5 VK: Well, Game of Thrones, that’s not a good example, because that was scheduled. Those didn’t all come out at once. So isn’t it if you’re not watching it at the schedule time it was released, isn’t that considered a rerun?

0:09:42.3 TW: No.

0:09:42.4 JH: Yeah, like if I record it and watch it, is that a rerun?

0:09:45.6 TW: No.

0:09:45.6 JH: I mean, they didn’t say, though. Is it a rerun to me, a re-won rerun to them? I can’t even say rerun anymore.

0:09:49.7 TW: We’re going to belabor the crap out of this, but literally a rerun is based on broadcast, is over-the-air TV, that they are rerunning something not at its original date.

[overlapping conversation]

0:10:03.0 TW: So I’m back to the like actually defining what a rerun is. I mean, there’s the stuff that is obviously Gilligan’s Island, The Office, those would, sure, that’s, it’s stopped production. It was produced years ago. But I do think you get into, what is their definition? And we’ve probably beaten this into the ground.

0:10:19.6 VK: Well, so who’s, what’s your favorite NCIS episode? Let’s go around the horn.

0:10:22.9 TW: Really?

0:10:24.5 JH: I don’t even know.

0:10:26.1 TW: Oh, you’re just trolling me.

0:10:27.0 VK: Yeah, just trolling. So get this, though. NCIS has, how do they say it, people have spent 11.4 million hours a week streaming NCIS since March of 2021, in other terms, roughly 11.4 million episodes a week. Is what they said. That was like further down in the article, of NCIS, supposedly, by their measurements.

0:10:52.3 MH: I’d like to think that’s like 10,000 people with no life at all.

0:10:58.1 JH: I used to definitely consume a lot of crime content as research for how I would evade my captor ’cause I’m so convinced that I’m going to get kidnapped someday. Because I’m telling you, you handle a narcissist very different than, Tim’s shaking his head. Than a sociopath, so I learned a lot.

0:11:17.8 MH: This is great content, but mostly for our sister podcast, True Crime with Val.

0:11:30.3 TW: We’re eight minutes into our first example.

0:11:30.6 MH: True crime situations with Val.

0:11:32.8 JH: All right, so then let’s put an end to this one. Go around the horn. I say it is poorly written and suspicious.

0:11:41.6 JH: Yeah.

0:11:42.8 VK: Would you agree?

0:11:43.7 JH: This myth is busted.

0:11:45.6 TW: I give it two stars. Can we change our ratings?

0:11:48.1 MH: Change our ratings? Yeah. I give it four Pinocchio’s, and… Yeah. No, I’d say I agree. It lacks a little context. And so what it’s trying to communicate to you is sort of getting a little bit lost in the shuffle. So that would be the takeaway I got. It sort of left me with more questions than answers.

[music]

0:12:12.6 MH: It’s time to step away from the show for a quick word about Piwik PRO. Tim, tell us about it.

0:12:18.8 TW: Well, Piwik PRO has really exploded in popularity and keeps adding new functionality.

0:12:24.4 MH: They sure have. They’ve got an easy-to-use interface, a full set of features with capabilities like custom reports, enhanced e-commerce tracking, and a customer data platform.

0:12:36.2 TW: We love running Piwik PRO’s free plan on the podcast website, but they also have a paid plan that adds scale and some additional features.

0:12:42.4 MH: Yeah, head over to piwik.pro and check them out for yourself. You can get started with their free plan. That’s piwik.pro. And now let’s get back to the show.

[music]

0:12:54.6 MH: All right, I’m gonna do one that I think is pretty decent actually. And in this one it says, nearly 5% of Americans don’t have a bank account per latest 2021 data. So that’s the headline. And then the story goes on to talk a little bit about while this is the lowest number of unbanked people, since they’ve tracked that, it still represents about 5 million people in the country who don’t have a bank account. So I thought that was actually a pretty good headline. It was not sensational. It was accurately stated what this data said and was pretty straightforward. I don’t know. Thoughts? Comments?

0:13:38.9 TW: Julie’s like, well, the format was supposed to be…

0:13:40.5 VK: Oh, so you mean like not the purpose of this episode?

0:13:45.2 MH: Oh, well, I’ve got others, Val. I’ve got others.

0:13:47.7 TW: And Julie’s like, so read the headline, get the reactions from everyone and then give your take on it. So, okay, cool.

0:13:53.6 MH: Oh, I just went ahead and gave my take on it right away.

0:13:56.0 JH: You just rolled right into it.

0:13:57.4 MH: Jumped the gun there.

0:13:58.4 TW: Yeah.

0:14:00.4 JH: I was gonna say, I mean, it sounds pretty solid. I have no reason to question it.

0:14:06.1 VK: I mean, I trust you, Michael.

0:14:09.4 JH: You said it so convincingly.

0:14:12.7 VK: Yeah. I guess if you asked me what percentage of Americans don’t have a bank account, I don’t know. It’s kind of like how many windows are in the city of Chicago kind of question. I don’t know how I would have backed into what my gut would have said around that. But I’d be interested to look at some of the source of how they track that, like the absence of something being present is always harder to track. So I’d be interested to look at the source of data.

0:14:40.3 TW: Yeah. I mean, the tough one, it’s so funny. It’s like a decent headline and we’re like, yeah, okay. Okay I can only do bad ones.

0:14:54.8 VK: Well, it’s like I learned something. I don’t know. It was like a new fact.

0:15:01.0 TW: What I mean, I suspect that is very, very low income people who are kind of paycheck to paycheck and those are the ones who use the cash. And then it’s the people who just don’t trust, which is probably a much, much smaller percentage. So I could see that if the article, like something more useful that might be a, this is an increase or a decrease in people who are living so close to the edge that bank account, like they’re living in the red and they’re paying to live in the red ’cause they’re having to use paycheck cashing services or whatever. So to Val’s point, the fact that it’s just a number that’s not necessarily anchored on something. To not belabor it, maybe I’ll counter this with one that I’ll just choose to segue to one that is more, the headline is, more teen girls smoke marijuana than boys now, study shows. I’m gonna go ahead and give the subheading was, overall, the percentage of teens reporting marijuana use fell to 15.8% in 2021 from 23% in 2011. And this is an article that came out in the fall of 2024. Reactions, headline, more teen girls smoke weed than boys, sub headline, the overall percentage in a 10 year period from 2021 to 2011 from 2011 to 2021 fell from 23% to 15.8%.

0:16:39.8 VK: And it was self-reported. Did I hear you say that in the headline?

0:16:42.5 TW: It’s not in the headline. The, it’s actually from an annual school-based self-administered online survey of US middle and high school students conducted from January to May of this year. Which is another odd bit of phrasing, but I’ll withhold my judgment.

0:16:58.7 VK: So they’re asking them 10 years ago if they did.

0:17:03.4 TW: Yeah, I think, well, it’s an annual survey. So I think they do it every year. Their phrasing referred to this year, but it’s reporting results of 2021 compared to 2011. And this was news that made the rounds in 2024.

0:17:23.3 JH: I love that you guys are like parsing the timeframes and data in my head. I’m like, Oh, I’m sure it’s just because there’s more people using edibles. They’re just smoking it less like that’s where my head went. That’s so funny, but.

0:17:36.8 TW: No, that’s, I mean, I’ve got, I got a lot of thoughts on this one, so.

0:17:40.3 MH: I mean, I think teen boys just lie to survey people, obviously.

0:17:45.7 TW: Yeah. I mean, right away, like the compare and contrast is sort of like raises my eyebrow a little bit, but yeah, I’d love to hear your thoughts.

0:17:54.3 MH: I mean, one, it seems like the lead would be that marijuana use dropped by 6%. But, on top of that, Val’s point, the consumption. So I went to vaping versus, I mean, edibles, another way to consume marijuana vaping. I’m pretty sure over that span of time has like skyrocketed and then has all sorts of bumpy things. The fact that they’re talking about this being a survey, like I know people, organizations close their books at the end of a quarter and then it takes them a few weeks to report results. I know there are macro economic indicators that it can take a few months to compile the data. If this is a fucking online survey, surely there’s not a three year lag and that they’re saying, Oh, we just got the numbers in for, because it actually says… So it says conducted from January to May of this year, but it’s also, they’re picking 2021. I don’t know if you guys remember 2020, 2021, kind of fucked up period. So 2021 in theory they were back, but I’m like surveying middle school and high school students in 2021.

0:19:09.0 MH: That is an atypical period. So it’s putting, it’s stating these things. I mean, it does say, a study says, so it’s not necessarily stating it as fact. It does give precision of a 10th of a percent in the subheading. But it just overall, I was like, the study may be, it’s a nice longitudinal study. Like if you do this every year, then… But you cherry picked two points 10 years apart. I’m like, oh, if you plotted that by year, is it like super noisy? And you said, what are the points where there was a big spike up and a big spike down and then report on that? Like, that seemed like it totally could be happening big time. So I just, the more that I thought about it, the more… And the fact is the difference between the girls and boys, like wasn’t that dramatic. It just like, that seems like the question I least have about this. And then everything within the article was just making me be like, eh, this feels like absolute cherry picking of data and not considering external factors like different consumption methods. A pandemic that went and with weird back to school stuff. So it yeah.

0:20:33.4 VK: You can’t smoke weed in the school bathroom if your school is your home, it’s a little harder.

0:20:41.1 MH: Yeah.

0:20:41.4 TW: It’s true.

0:20:41.7 MH: I mean, they said 2020, if they’d done 2020, I would’ve been like, come on. But 2021, I was like, that was… I mean, schools were back in. But I mean, I know from separates, I mean, there was huge, well massive swaths of kids. Like never showed back up. I’m like, huh, I would wonder if the people who once they had that year of remote schooling didn’t return, if maybe those would be in environments that would skew more heavily to smoking weed, maybe. I mean that’s a theory I have no… I don’t have evidence on that. It just seemed like a very, very surface level kind of making statements. So.

0:21:24.3 JH: Definitely.

0:21:25.5 MH: There is my thoughts on that one. I think that means, Val, you’re up.

0:21:29.0 VK: Okay. Well I’m gonna try my best to stay focused on this one. All right. So here is a headline that was, Why don’t women use Artificial Intelligence?

0:21:43.7 MH: Oh, oh wow. Yeah. That’s…

0:21:49.5 VK: Just.

0:21:50.2 MH: I’m.

0:21:51.0 JH: Kroll just blanket statement. Don’t use it at all. Like. I’m pretty sure we could attest that at least I, myself have.

0:21:57.1 MH: Why don’t girls play video games?

0:22:00.0 JH: Yeah.

0:22:00.0 MH: Like, that’s a terrible, I mean, wow. Interesting.

0:22:03.9 JH: Wow.

0:22:04.0 MH: Without knowing anything about the article, I would be like, what kind of question is that?

0:22:10.7 VK: It was published in The Economist.

0:22:12.4 MH: Oh, well there you go.

0:22:15.6 TW: So that’s Yeah. That’s Fly By Night Rag.

0:22:18.7 MH: Yeah.

0:22:20.6 TW: That’s…

0:22:20.6 MH: Yeah, that’s hard. Like, okay, so what were they trying to say there?

0:22:27.1 VK: I don’t know. It feels Clickbaity too. ‘Cause they’re like, just pose the question hoping you’ll click and read instead of telling you.

0:22:33.6 MH: So I feel like I saw some posts that I’m assuming were grounded in this, that the underlying was that some attempt to measure the rate and it was showing that a lower percentage of women were using AI tools than men. And so this was a gross generalization. I have issues with some of the commentary on that as well, but ’cause yeah, it’s clearly women, like you just classified half of the population and said they don’t do something. So it’s factually cannot be right. And it definitely looks like clickbait, which seems weird for the Economist, but.

0:23:18.2 JH: Agreed. Any other thoughts before I share some of my.

0:23:23.5 VK: Give it to us.

0:23:24.3 JH: Well, so I actually went to find the paper that was being referenced and that the paper is called Global Evidence on Gender Gaps and Generative AI. And so there’s like Harvard Business School, Berkeley, Stanford all got their names all over this. And basically it was a meta-analysis of 13 different studies. And none of these studies were aimed at looking at the gender gap between adoption. It just happened to be one of the different factors that was captured along the very, I mean, the studies, some of them were in Kenya, some of ’em were in Spain. They were literally all over the globe about AI adoption. And so none of them were trying to demystify this. This was just a meta-analysis of them. But 26 countries were included. I mean, even where we don’t even have to scroll down too far before they say, but there were exceptions. BCG had a study showing that women are 3% more likely to use AI than men in a sample of nearly 7000 US Tech employees. And so then it has like all these concessions. So it’s basically saying and a lot of these, we were able to find this even though that wasn’t the intention of the research, meaning that it wasn’t even necessarily controlled for. And we also found some that weren’t conveniently aligned with this headline. But we’re gonna roll with that one.

0:24:38.9 VK: Anyways, just gonna stick with the storyline. Wow.

0:24:42.3 MH: Just.

0:24:43.1 VK: That’s pretty brilliant.

0:24:45.4 MH: Stay away from that AI ladies, please.

0:24:47.9 JH: Yeah.

0:24:48.1 MH: Don’t even touch it.

0:24:49.4 JH: And there was actually, there was…

0:24:50.8 TW: Well there was.

0:24:51.1 JH: Go ahead.

0:24:51.5 TW: But also, another one you caught in there was like, it says, why don’t women use AI? Which or why don’t women use artificial intelligence? And then the study immediately went to generative AI, which is absolutely…

0:25:09.2 JH: Oh, yeah. Didn’t even catch that.

0:25:09.3 TW: A small subset not to mention defining use. Yeah.

0:25:16.2 MH: Yeah…

0:25:18.5 VK: But there was another sentiment in here, which I really loved, that was asking, you know, why wouldn’t women want to use generative AI to advise? Because it can help them automate tasks to help manage work and family time demands, which I love.

0:25:34.9 MH: Wow. So some of the people meta-analysis were from the dudes from the 1950s?

0:25:41.7 VK: Yeah, apparently men have figured out all the use cases for how to handle the burdens of their home.

0:25:46.5 MH: Listen, a little lady, you can have a look up the recipe for a perfect martini for me when I get home from work, ’cause it’s been a hard day doing my prompt engineering.

0:25:55.3 TW: I feel like most, or any commentary at this point is very dangerous for me.

0:26:05.3 JH: It was, I mean, there were some points in there, just, it was trying to make the point that like, just giving people access to AI isn’t gonna get them to adopt the technology, which like, I think we all could have come to that conclusion before. But a couple of these studies were just about, you know, WhatsApp for business released an AI function, like a feature within the tool, and fewer women happened to click on it. I’m like, so this is about, we’re painting some brush, broad strokes about our ability to adopt or use our leveraged software just ’cause we didn’t happen to click on a new feature. So anyways, it seemed like quite the jump.

0:26:40.2 TW: Yeah, but that one does seem like the applicability to, I mean, you broaden it to where, oh, there’s just something collected on as a matter of course, on the various ways we do lead forms. And then somebody gets asked to do, you know, maybe a junior analyst, or maybe even just, you know, a marketer says, oh, I’m gonna do this comparison, and oh, look, this is 55%. And this is 44%.

0:27:06.4 TW: And they, they jumped to the massive thing that is missing there is that you kinda hit it on at the very beginning that ’cause that wasn’t the research question being asked. And now you’re trying to repurpose this data, it just like immediately should trigger a million red flags. And then by the time it gets filtered through to the headline writer, like it’s just stating something like a rhetorical, it’s not a rhetorical question, it’s just like a, it’s stating… It’s implying a fact that is so far from a fact. And it’s so vague that it’s like horribly irresponsible. So I thought it’s a good one.

0:27:51.3 VK: And just to belabor the point a little bit, Tim, but it is interesting drawing the parallels of like, when people have an actual designed experiment, study methodology, whatever, and then your stakeholders ask you to, like, get other insights than the main one it was made to get and people are so, like, you know, loose about the idea of, oh, well, if I designed the study, then everything about it in comparison, like, is fair game. And we know that’s not true. But I do think people obviously, obviously fall into that, that trap.

0:28:24.6 TW: Which seems like, I mean, that would be the sort of thing that if there was… It seems like things happen around gender certainly. I think stuff happens with kids or different backgrounds or different things where in the social sciences, they are used to saying we should field research to answer this. And if we field a well-designed study, it does not have to be horribly cost prohibitive, but can give us a pretty solid answer without doing necessarily like a field experiment.

0:28:56.3 TW: Although that sort of research is only gonna give you the what is happening, not the why. Which is maybe the other part is like, this is asking for causality, and I’m sure there’s like speculative garbage in the article, but those are all like, one, you’re starting with a premise that is so flawed, and then you’re trying to find, speculate on the causality about something that who the hell knows if it’s true or not. Wow. Okay. Well, cool. Triggered. Well done.

0:29:27.0 MH: It seems like, yeah, definitely the article could have had a much different headline and been very useful. Like, is there a gender gap in generative AI usage or something like that? And just ask the question, and it’s a very intriguing question. And I think a lot of people would wanna dig in and understand that better, but it’s sort of like, why don’t women use AI? It’s like…

0:29:50.3 JH: Goddammit.

0:29:53.0 TW: It’s like, we built it just for them. Like, come on.

0:29:57.0 JH: Help them manage their households.

0:30:03.9 JH: Oh, man.

0:30:05.3 MH: Oh.

0:30:07.0 MH: Ooh, that’s a good one.

0:30:07.9 VK: All right. I’m gonna pass the stick back to Julie.

0:30:14.1 JH: All right, me. Okay. Next headline. The order in which you acquire diseases could affect your life expectancy. New research.

0:30:29.1 VK: Well, I’ll make sure that I acquire them in the right order after reading this.

0:30:29.2 MH: It’s always the last one you acquire that kills you.

[laughter]

0:30:33.3 JH: Oh, damn. It got me. You know?

0:30:37.6 TW: Should have got this one first. Darn it.

0:30:40.5 JH: Yes.

0:30:41.2 MH: And it’s like, the thing you’re looking for is always in the last place you look. Oh, my gosh. Sorry. That’s funny.

0:30:54.4 TW: What is the underlying idea? I mean, yeah. The earlier you get a really serious life-threatening the disease, the more likely it is to kill you. Right, Tyson?

0:31:05.4 MH: In order.

0:31:06.0 VK: Or maybe it’s like how you can get chicken pox when you’re younger, and like the virus is always in you, and that’s what shingles is, and things like that. So if you didn’t get chicken pox, so it’s not like you’re gonna get shingles first when you’re six. So the order bit of that is what’s interesting to me, ’cause there’s so much now about customized medicine.

0:31:27.3 TW: Yeah, you should get adult-onset diabetes before you’re 12.

0:31:30.0 MH: Yeah.

0:31:30.1 VK: We’ll have so much time to try to handle that.

[laughter]

0:31:35.2 TW: I’m just, the thing going through my head right now is multi-disease attribution. It’s like, which disease was influential, and which was really close?

0:31:48.7 VK: Well, first touch, last touch.

0:31:50.7 TW: Yeah, the first touch.

0:31:51.2 JH: Yeah, attribution models for disease.

0:31:53.1 TW: Multi-touch disease attribution. Oh man.

0:31:57.9 JH: So yeah. Yeah, you all mirrored my first reaction pretty well. So reading this article, it gives a lot more detail of why I think our reactions are valid. So it started off saying pretty much just in general, it said like, using statistical models, we examined order and timing of developing psychosis, diabetes and congestive heart failure in patients of the same age, sex and area, deprivation and the related impact on their life expectancy. So that was like the first explanation piece I could find. First thing I thought was why those three things and tell me those three things. And they didn’t really mention those three things in the headline. I’m like, that’s a really interesting pick. But yeah, those sound pretty serious. So not just diseases like ends up, they were only looking at, how do they phrase it?

0:32:49.7 JH: They were only looking at multiple long-term conditions. So not, not just diseases, long-term conditions. And the only three they looked at were the three I listed. So then they ended up looking at it, they did say that they had a pretty like robust data set, but even so, and it was like over 20 years and whatnot. But even so, they ended up looking at the different combinations of just having one, just having two, having all three and then in what order as well. And it was kinda crazy. They said, oh, and they didn’t mention the exact method either. I did try to find that and they never listed that anywhere. But pretty much the results were saying that congestive heart failure is the one that cuts your life expectancy the most. Which I don’t know if I’m surprised by that. If in the list of three options.

0:33:38.3 TW: That’s a headline right there.

0:33:40.7 JH: And they said, yeah. So if you get that one third, that cuts the most life off of your, expectancy.

0:33:48.8 TW: It’s always the last one that gets you.

0:33:51.8 JH: Yeah. Yeah. The last one. That’s what I thought. I was like, it’s always the last one that gets you and the most acute, I would say maybe sounded like congestive heart failure.

0:34:01.9 MH: It was congestive heart failure, psychosis. The what was…

0:34:04.2 VK: Diabetes?

0:34:04.7 MH: Diabetes.

0:34:05.6 TW: And diabetes.

0:34:05.7 JH: Yeah. The one surprising result that they did outline was that you have a longer life expectancy if you have diabetes and psychosis rather than just psychosis. And they hypothesize that it’s because if you have diabetes, you’re seeing healthcare more regularly.

0:34:22.8 MH: Yeah.

0:34:22.8 JH: Which might actually help you better manage it. So I was like, oh that is actually really interesting. Never would’ve thought that, but the big bad congestive congestive heart failure. Yeah. That was, I felt a little obvious.

0:34:36.4 TW: So counterintuitively, if you are have psychosis, you should just eat a lot of sugar so that you get diabetes, ’cause then you would, so anyway, I don’t mean to make light of some of these illnesses. That’s not what I mean to do, but…

0:34:49.6 JH: Yeah.

0:34:50.3 TW: Yeah.

0:34:51.0 JH: But it was interesting, they talked a lot about the order and in the end they were saying, oh, it doesn’t matter the order. But they did find more about the combination. But I still am just wondering why they chose the heart failure piece and like that combination of three. And by the end of the article, what I did find, not reassuring but like a better thing to focus on was the fact that they were saying this method that they analyzed this data with, could be applied to different diseases, different combinations of things. And it could be really helpful for different programming and knowing if someone gets one disorder, should you be on the lookout for another specifically. And I was like, oh, that seems way more helpful than like calling out necessarily the top three. And then again, I thought the headline was a little misleading once I got into it.

0:35:39.6 MH: So that they’ve got, if you take the sequencing out of it and you take the combinations of each single, each had two, each had three. I also like what is the scale? Like what is the actual broader occurrence in the world? Like if you get it down to where, yeah, this tiny little sliver, it’s really, really bad. But it’s a tiny little sliver of people. Like I just… There’s part of me wonders if there is some sampling error risk, like finding ones who only had one was presumably easier than finding people who had two, which was easier than finding people who had three. And then you put the sequencing for two or three. That seems like you inherently would wind up with pretty small numbers that you’re then trying to draw inferences about the whole population. And somewhere it feels like there might be some oddly survivorship bias risk, but I can’t quite frame what it is. Like if you had a couple of these and that meant that you actually were so much more likely to die, that you actually were never like eligible to be found in the study. ‘Cause they were only looking at people who were 18 to 45 or something. I don’t know. But when you’re trying to look at those combinations of factors and you’re dealing with a sample, things get pretty dicey. ‘Cause you wanna make sure your sample is actually reflective of the population.

0:37:18.8 MH: Well then what happened to cancer and Alzheimer’s and…

0:37:23.1 JH: Yeah. Like did they look at people that had other ones? That’s a good point too.

0:37:27.6 MH: It seems like they just sort of like, well we’ve got data for these three so.

0:37:30.2 VK: Exactly.

0:37:30.6 MH: We are gonna just do this analysis.

0:37:31.7 VK: Well, that’s what it felt like to me. Like after 20 years we have to find something to report on. Like we, there’s gotta be a headline in here somewhere.

0:37:42.1 MH: Yeah. As an analyst. I think we’ve all sort of felt that pain before we’ve got us report something back.

0:37:49.3 JH: With all those data.

0:37:54.2 TW: Your marketing campaign has cancer.

0:37:56.9 MH: Yeah, that’s a good one. Corporate popcorn.

0:37:58.7 VK: Yeah, corporate. We’ll go with Tim.

0:38:03.1 TW: Oh, okay. So this one may be a little bit of a cheat. Cause I found it ’cause it had come out that somebody had like busted it.

0:38:12.8 TW: It’s not quite a headline, but a few years ago it really made the rounds in the media and it was became a Netflix, Netflix documentary called Live to 100: Secrets of the Blue Zones. You familiar with that?

0:38:34.8 MH: I’m literally planning to move to Loma Linda, California. I think it’s a blue zone.

0:38:36.9 TW: Is it a blue zone? I thought the blue zones were all like…

0:38:40.9 MH: Oh, there’s one like in the middle.

0:38:42.7 VK: There is in California.

0:38:45.6 TW: There’s one in the… Okay. Thoughts on that.

0:38:48.4 JH: It wasn’t a headline and I don’t know, I’m not familiar. So no thoughts.

0:38:51.7 MH: I mean, so the blue zones were places that people were live wildly longer. And so then it became a whole like, well, let’s figure out what’s unique about those different countries and regions. And therefore we will find the secret to longevity.

0:39:10.2 TW: Yeah. I mean, I thought they were real, so I just started eating a lot more olive oil and feta cheese. Not really.

0:39:21.8 JH: Mediterranean diet.

0:39:21.9 TW: Yeah. Just Kalamata olives, that sort of thing. Yeah.

0:39:27.3 JH: I mean, I think it goes a little bit to what we were just talking. Like my first reaction is I wanna believe it’s true as well, ’cause you’re like, how are large numbers of each population that they looked at like living that long? But at the same time, like, I don’t know. How do you really control for factors to prove some of the things they’re observing are the cause of it? Like, I don’t know.

0:39:47.9 JH: Or is it, how do they know it’s not a combination of a bunch of things? Some of the things you’re not aware of, you know, and it’s just the one thing they kind of focus in on the show, like the walk, I think it was an Italian village where the walking, they like really honed in on walking. And I think in Japan, it was similar. It was like sitting down and getting up even at a hundred, like they sit on the floor.

0:40:08.4 JH: And obviously some of their diet there, but I guess those are my first questions I would have.

0:40:15.0 MH: So no one has gone to, maybe it’s just bad data. So there’s this guy named Justin Newman won an Ig Nobel this year because he started digging into what those blue zones were.

0:40:31.4 MH: And he posited that, oh, a lot of these are in these regions where like record keeping is shitty. So there are lots of people who show up as being old. And then if you’re getting any sort of check, ’cause grandpa lives with you and you’re getting some from your government and they die, like.

0:40:56.3 MH: Do you wanna report that they’re gone or do you want that to keep coming in? So it’s actually like late age and death record keeping, is actually pretty hard, like, and people will kind of keep getting a pension check. So basically a confounding variable that it’s actually really tough to track when people die. Having said that, I mean, so people went straight to applying, causation.

0:41:28.7 MH: What are the things is it locking, walking long distances, eating this, is it having familial connections? I kind of wonder, even the, this Justin Newman’s like study, I think raised like, oh, these are like some good questions. People just accepted his fact and he has some serious questions about the data. I’m not a 100% sure that his counter is entirely valid either, I think a lot of things that were found, people were like, yeah, if you move more, you’ll live longer.

0:41:57.9 MH: So there was a degree of, there were causal factors, but it was one of those where you could find enough things that explain, and then you could kind of tell a story that started to completely make sense, but I’m not sure. I’m not sure I entirely buy into the debunking, but I also think it is pretty funny that entire like companies and books have been written and they just like accepted his fact, and I think he did find some pretty definitive records of saying, yeah, people aren’t living to be that old. That’s just… You’re just looking at the government recordkeeping in some of these countries, that’s, it’s just not that great and they’re fine with it. So there you go.

0:42:41.3 JH: I was not expecting that one.

0:42:43.6 VK: I thought it was some like death becomes her kind of situation.

0:42:47.8 MH: Yeah. What a downer. Oh geez.

0:42:53.5 TW: Michael, you got another one?

0:42:56.1 MH: I do. All right, here you go. This time I’ll wait and get reactions. Here we go.

0:43:00.7 TW: Do it right this time.

0:43:07.2 MH: Here’s the headline. Thousands of cleaning supplies may contain substances linked to health problems.

0:43:14.4 JH: Well, yeah. They’re literally full of chemicals.

0:43:20.0 VK: The poison control, like Mr. Yuck stickers over everything under my kitchen cabinet.

0:43:27.8 MH: It’s like thousands.

0:43:29.1 VK: I mean, I was planning on drinking all of it, so I’m really thankful that this headline’s really cut me off at the pass.

0:43:33.8 JH: Could be harmful. Hmm. Who’d have thought?

0:43:40.8 MH: Wow.

0:43:40.9 VK: Where was that published?

0:43:41.7 MH: That was from CNN. So yeah, while looking for source material, I was like, USA Today and CNN are good areas for me to wander around and try to find bad headlines. So yeah, I feel like that one is sort of like, no duh.

0:44:03.2 MH: Like if you presented a headline like that in a business meeting, people would be like, yeah. It’s sort of like, the more sales we make, the more revenue we get.

0:44:14.2 MH: Wow You’re never invited to this meeting ever again Yeah, so, try to avoid that but it’s interesting ’cause in the article they basically were like, yeah, these are dangerous and you should be careful with them But don’t go throwing away all your cleaning supplies ’cause that’s bad for the environment to waste them. Just read the labels.

0:44:36.9 TW: And don’t not clean your house because…

0:44:39.9 MH: Yeah, no you should clean more it’s a weird piece, honestly.

0:44:43.7 JH: Spray those chemicals everywhere.

0:44:48.3 MH: It’s like so what am I supposed to do? I’m just supposed to freak out about how the chemicals might be hurting me, but don’t do anything about it, and then read all the labels very carefully and then try to educate yourself on the top 500 terrible chemicals that might be in a cleaning product. It’s like really There’s a website they talk about that have like a safe use like I think the EPA has like this is safe to use registry or something that people can apply to so if you don’t know you can look that up on your phone in that grocery store.

0:45:23.1 TW: But I mean it does it does bring out. I mean, it’s funny to like map that to a marketing context like they’re their trade-offs like do you wanna not clean stuff.

0:45:31.2 TW: Do you wanna clean stuff? But it’s like takes forever ’cause it’s like it’s a horribly tedious ’cause it’s totally organic and it only does 70% of good of the job or do you wanna clean it like really effectively really Quickly like those are there different benefits so there are trade-offs and it is very easy to say let’s paint the… State something that is fairly obvious, but I can imagine in a business context where you you find something that is short term oriented, doesn’t actually consider the counter side of it and it’s kind of hoped like oh, I found this thing. But that’s just a slice of the consideration.

0:46:19.1 JH: But also I guess that’s like now that I’m thinking about a little more digesting the initial shock of the headline. Is I think also, in what way is it dangerous like obviously any chemical in a potent enough form isn’t good for you But like is it don’t spray it on your tongue. Don’t inhale the mist. Don’t touch it 10 minutes after, don’t touch it at all.

0:46:38.7 JH: Like what in what form or in what contact is it really like this is dangerous for you but it would be good to use on a counter or something really dirty right and then you wipe it away and 10 minutes later. It’s like done its job and it’s neutralized. Like I don’t know I you start to get a lot of follow up questions.

0:46:58.2 TW: There’s thousands of cleaning supplies Julie and they may be dangerous.

0:47:00.4 JH: Maybe.

0:47:00.5 VK: So Watch out.

0:47:00.8 TW: Watch out.

0:47:00.9 VK: I mean as someone who grew up in a household that every Sunday smelled like bleach comet pine salt and pledge, when my husband and my daughter was born went through a phase of like well, let’s clean everything with with what is it? Red wine vinegar or white vinegar? I was like…

0:47:18.5 JH: Oh, yeah.

0:47:19.7 VK: Hard pass, absolutely not absolutely So bring on the harsh cleaning supplies I’ll deal with those health problems ’cause they’re gonna be in the right sequence and I’ll be fine.

0:47:31.0 JH: That’s right.

0:47:32.0 TW: You also have to not mix the wrong ones together to the article did talk about that.

[overlapping conversation]

0:47:36.6 MH: Right bleach and ammonia right makes the…

0:47:42.0 TW: Bleach and ammonia.

0:47:42.3 MH: And that was because that was so bad when I was in college We would we had to clean the grease traps in the kitchen like twice a year. And so we would take them we would go outside ’cause we were gonna be responsible. We’re like wow if you put those together and they’re like brutally toxic like that probably cleans the shit out of the grease so we would definitely take to like big garbage buckets put hot water into it and then just go all in with the ammonia and bleach and kind of stand back.

0:48:05.5 MH: We had gas masks. I think tube or that we’re outside. We’re fine.

[overlapping conversation]

0:48:08.4 JH: Oh my gosh.

0:48:09.3 TW: So what he might have been folks.

0:48:21.3 JH: Here he is, read headlines on the podcast.

0:48:25.8 TW: That’s right. Wow.

0:48:27.6 MH: That could have been somebody.

0:48:29.2 TW: I mean I’ll admit to being curious about it, but I never was like hey, let’s do that ’cause like I always got the warning from my parents like never mix these two chemicals.

0:48:34.3 MH: Yeah, we were definitely more of the like, oh, come on.

0:48:39.5 TW: You’re gonna die if you mix these two chemicals, okay.

0:48:42.0 MH: Yeah, I mean the fact that we had gas masks to clean the kitchen twice a year, I mean there was lots of other stuff going on. That was…

0:48:47.4 VK: Yeah, I had them at the ready.

0:48:51.5 MH: Yeah. They were good ones. They were full-on.

0:48:53.4 VK: Army surplus.

0:48:56.4 MH: I Think the other thing that stands out is like this is a needlessly sort of like a scary headline to you. It’s sort of like alcohol website is broken and no one can use it, right? It’s like okay, that’s not helpful at all, right? You’re not really getting us anywhere.

0:49:10.7 VK: There’s a broken link in the footer, yeah.

0:49:17.5 MH: Okay, specifics…

0:49:18.8 TW: Or shout out to Ezekiel. Yes, the social link posts on our website still do not link to the specific episode post can’t figure that out, but it’s a known bug a known bug one of many known bugs on our website.

0:49:37.2 MH: Still less bugs to GA for. All right. All right, Val, do you have another one we can sneak another one in.

0:49:42.9 VK: Sure. So when I read this one, it just felt very suspicious I’ll say so this one came out earlier this year it says Google delays third-party cookie deprecation likely until 2025, I’ll get your thoughts.

[overlapping conversation]

0:50:05.0 MH: I always wanted to talk about third-party cookies.

0:50:09.6 TW: I thought they were they gave up on deleting those all together…

0:50:13.8 VK: That’s why the headline sounds kind of unbelievable, doesn’t it?

0:50:16.9 MH: Okay, so I was like, was I misinformed.

0:50:19.8 JH: Well, it sounds like assertive in the beginning and then it throws in the likely and it kind of threw me.

0:50:25.9 VK: Yeah, that was a throwaway. Okay. Sorry. We don’t pretend like that was real. Okay So this is actually an older one and I loved using this in presentations back in the day trying to explain some of the benefits of being thorough, but the headline is too much running is tied to shorter lifespan studies find. Have you guys come across this one?

0:50:49.4 MH: No.

[overlapping conversation]

0:50:53.7 TW: Crap.

0:50:54.5 JH: What?

0:50:54.6 TW: Always the last mile you run that kills you. Bro if you’re running it with congestive heart failure, followed by psychosis. That’s crude.

0:51:05.9 JH: Not in a good place.

[chuckle]

0:51:08.0 MH: Too much running.

0:51:09.5 VK: And there’s been lots of articles about this one. So there’s another that’s on Runner’s World. This one is, “Running too much will kill you.”

0:51:15.7 TW: Yeah. Having read that article a long time ago, I have really put that into practice, so I’ll tell you that right now.

[laughter]

0:51:23.9 JH: Avoid it. I just avoid it.

0:51:25.6 TW: I plan to live for a very long time.

[laughter]

0:51:28.8 MH: This seems like the ones that are like, “You can be… You can over-hydrate. Drinking too much while exercising can cause a problem.” But I’m trying to… This should be super obvious to me as to why this is factually accurate but absolute garbage, and I’m missing it.

0:51:48.9 VK: So it’s more obvious than you think it was. So it’s… The study was low sample sizes, first of all, and it was broken into three groups of low running or no running, moderate, high, and then there was the age. And so the group that were older and running at a high amount, whatever the mileage was, like 20 plus miles per week or something, there was only three people in it. And over the duration of this 10-year study, they died. So it was like mortality rates, they all died.

0:52:18.8 JH: Oh, jeez.

[laughter]

0:52:21.4 VK: But it was the oldest group, only three people, so yeah, it was just this… It was…

0:52:26.0 JH: Jeez.

0:52:26.1 MH: I wonder if there’s anything wrong with our causal model. No, I don’t think so.

0:52:31.1 JH: Crazy. Well, yeah, ’cause my first question was gonna be like, “How are they defining too much in the study? Like, how did they even look at, yeah, how much are you running and then control for all those other things? And then we’re gonna follow… ” I mean, how long was this study to know that your lifespan was cut short? Who did they control against? Like, that’s crazy.

0:52:50.4 VK: All the questions. But if you look for… It didn’t take too much work just for me to put my hands on this one. It was like, “Too much running will kill you.” And I said like, “Circa 2014.” And it was like… You’ll find so many articles of people taking this and running with it, no pun intended, or breaking down why this is such bad research and not a good steward of research practices or data headlines.

0:53:17.8 MH: But there is a historical context here, ’cause the guy who ran the very first marathon apparently collapsed and died right afterwards, right?

0:53:26.4 TW: Right.

0:53:26.6 TW: Back in was it…

0:53:28.1 MH: That was too much.

0:53:29.3 TW: Pheidippides or whatever? I don’t know his name.

0:53:31.1 MH: So the whole distance running movement is actually just a population control design. They’re like, “This is a good way to… ”

[laughter]

0:53:37.8 TW: “It’s basically a suicide cult.”

0:53:40.7 MH: Yeah.

[laughter]

0:53:42.5 VK: Heaven’s gate.

[laughter]

0:53:47.6 MH: Did we learn nothing from the ancient Greeks? Like…

0:53:54.3 MH: Wow. They’re terrible. That’s good. All right. Well, you, as you’ve been listening, have probably thought of a few headlines you’ve seen over the years that have been kind of like, “What’s going on here with this data?” If you wanna share a couple, we would love to hear from you. You can reach out to us at LinkedIn or on the Measure Slack chat group, or also by email at contact@analyticshour.io. And we’d love to hear from you. And I think Julie, Val, Tim, thanks for doing this. It’s been a lot of fun. And it’s fun to dig in and look at data from another perspective.

0:54:36.4 TW: Fun for us.

0:54:37.0 MH: Yeah. It’s just a little mix-up. We did a little switch-up. There’s probably a study out there that says listening to too many podcasts is gonna kill you also. But until that study comes out, we’re gonna keep doing podcasts. I don’t know. Any closing thoughts, anybody?

0:54:53.7 VK: Check your sources. Yeah.

0:54:55.3 MH: Check your sources?

0:54:56.2 VK: Question everything.

0:54:57.7 TW: I mean, I will say the… I thought this was gonna be a lot easier to find stuff than it turned out to be. So that actually was…

0:55:07.1 MH: Yeah, I agree with that.

0:55:08.3 TW: Somewhat glimmer of a positive like, “Oh.” I think there is… A lot of these big media, they have data journalists. They have been banned in the past. They have become laughingstocks for one reason or another. We should call out that we explicitly said we weren’t doing any kind of political agenda oriented stuff. So putting those aside, reasonably, responsible journalism may put out stuff… Like, the ones that we found that were bad were kind of… Yeah, they just… They needed some filler, and they found something and stated something silly. But this wasn’t… I feel like this would have been… Was much more egregious 10 or 15 years ago. And there is an increase in responsibility and sophistication inside publications in journalism.

0:55:58.7 MH: I took the other thing, Tim. I was like, “They’re not even bothering to share good data anymore.” That’s what I was…

0:56:08.7 MH: I was looking for that article. Like, you could find studies like, “Study shows drinking coffee increases your lifespan.” And then another one is like, “Study show drinking coffee reduces your lifespan.” It’s like, “All right, who’s writing these articles?”

0:56:21.3 TW: Well, it seems like it used to be that some of these media outlets would just conduct their own research, just completely irresponsibly go out and ask 20 people on the street and they would state their conclusion. Now they’re generally trying to cite studies. They do caveat them. They do talk to somebody who can speak knowledgeably about their limitations. Headlines are tough.

0:56:45.2 JH: Yeah. But hopefully, this inspires you for the internal headlines you get at work. Maybe question those a little more.

[chuckle]

0:56:52.1 TW: Oh, very nice. I love it.

0:56:52.3 MH: Look at Julie bringing it home. Like, this is why you listen to this whole thing.

0:56:56.7 TW: Full circle moment.

[laughter]

0:57:00.0 VK: You made it this far.

0:57:00.9 JH: If you put up with this till the end.

[laughter]

0:57:02.0 MH: 99% of podcast hosts on the Analytics Power Hour think you should take… Pay more attention to your titles.

0:57:09.6 JH: Just my arm doesn’t agree. Just that point, that 1%.

0:57:14.3 TW: Well, Mo wasn’t here today, so we don’t know what her perspective is. Might be totally different.

0:57:17.0 VK: She’s only 1% out of five hosts. Yikes! Dig into that data.

0:57:26.7 MH: Frankly, I just need to understand whether she’s using AI or not, to be really honest with you. Before we can really, you know.

0:57:33.0 MH: All right, listen, this has been fun. It’s been a great time hanging out, talking about this stuff. And of course, no show would be complete without a huge thank you to Josh Crowhurst.

0:57:43.9 MH: Our producer does all the stuff behind the scenes to make this show possible. And you can catch us on where I said before, and also we’ve got a YouTube channel now. So check that out, subscribe, do the things that you do on YouTube.

0:57:57.4 MH: It’s a good place to see the show happen too. All right, well, I think I speak for all my co-hosts here. Val, Julie, and Tim.

0:58:05.5 MH: When I say, no matter what the headline is, keep analyzing.

0:58:10.6 Announcer: Thanks for listening. Let’s keep the conversation going with your comments, suggestions, and questions on Twitter at, @analyticshour, on the web @analyticshour.io, our LinkedIn group, and the MeasureChat Slack group.

0:58:26.6 Announcer: Music for the podcast by Josh Crowhurst.

0:58:34.2 Charles Barkley: So smart guys wanted to fit in. So they made up a term called analytics. Analytics don’t work.

0:58:37.4 S5: Do the analytics say go for it no matter who’s going for it? So if you and I were on the field, the analytics say go for it. It’s the stupidest, laziest, lamest thing I’ve ever heard for reasoning in competition.

0:58:49.0 MH: All right, I’ll give us a five count. We’ll get this thing rolling.

0:58:56.6 TW: Of course, this is right when my dogs decide it’s time to bark and stuff. I don’t know if you can hear that or not.

0:59:00.9 VK: It’s squirrel o’clock, huh?

0:59:02.0 TW: Yeah, exactly. They’re ridiculous.

0:59:07.9 MH: Try to do a podcast.

0:59:13.4 MH: We can just maybe go around. Let’s share one and we can, we can talk about it. I think it’ll be, it’ll be kind of fun.

0:59:15.5 VK: Sounds good.

0:59:19.1 MH: Who wants to start? I’m just running exactly like we planned and scripted it out.

0:59:22.9 TW: Yeah, just stick to the script.

0:59:25.5 MH: Yeah, but I’m trying to give the impression of spontaneity, Tim.

0:59:31.7 TW: Oh.

0:59:31.8 JH: It’s quite a fragile balance.

0:59:32.4 MH: Six out of seven podcast hosts agree.

0:59:37.5 JH: That howdy doody is definitely how you wanna start an episode.

0:59:40.8 VK: It’s a fan favorite.

0:59:46.2 MH: It’s a way to start and, you know.

0:59:49.4 JH: Not wrong.

0:59:50.5 VK: No, you’re not.

0:59:54.1 TW: Maybe that’s what we do, though. Whoever does it gets to pick that. It’s like when you pass the…

0:59:54.2 VK: Talking stick?

0:59:55.3 TW: The talking stick.

0:59:57.9 JH: What do they call it in corporate popcorn? That’s why I yelled popcorn. And then I said…

[overlapping conversation]

1:00:09.1 VK: You’ve never heard that? I’ve been on a…

[overlapping conversation]

1:00:11.4 VK: Let’s do popcorn.

1:00:11.4 MH: I thought this is what we bought from the Boy Scouts every Christmas. So it’s a…

1:00:15.6 MH: It’s corporate popcorn.

1:00:15.7 TW: That’s how it winds up in the office ’cause you can’t take it in.

[overlapping conversation]

1:00:15.7 MH: Three different flavors, everybody. Help yourself.

1:00:27.9 VK: He does.

1:00:28.2 MH: Oh.

1:00:28.3 TW: Rock flag and eagles live 53% longer than squirrels.

1:00:41.3 MH: It’s ’cause the squirrels are getting killed by the eagles.

1:00:42.4 TW: By the eagles.

1:00:43.0 MH: Yeah. And I have this hasty Googling that is absolutely a garbage number, so.

1:00:50.5 JH: You Googled it?

1:00:51.6 VK: That’s more responsible. I thought you really just made it up.

1:00:53.2 MH: It was essentially what turns out. There are lots of different types of eagles, lots of different types of squirrels, and lots of broad ranges for all of those. So I eventually said, yeah, I’ll make it up.

1:01:01.9 JH: Tim’s so responsible. He couldn’t even make it up for a rock flag.

1:01:12.6 VK: Meta analysis.

1:01:15.4 MH: Quit, yeah. A meta analysis, a distribution of eagles versus squirrels on a histogram, like.

1:01:20.1 TW: I’ll get back to you on that.

1:01:21.5 MH: I think we got a super weak time.

Leave a Reply



This site uses Akismet to reduce spam. Learn how your comment data is processed.

Have an Idea for an Upcoming Episode?

Recent Episodes

#260: Once Upon a Data Story with Duncan Clark

#260: Once Upon a Data Story with Duncan Clark

https://media.blubrry.com/the_digital_analytics_power/traffic.libsyn.com/analyticshour/APH_-_Episode_260_-_Once_Upon_a_Data_Story_with_Duncan_Clark.mp3Podcast: Download | EmbedSubscribe: RSSTweetShareShareEmail0 Shares