Java Jive: A Second Cup1
Check out the following newspaper headlines:
Drinking Coffee Can Help You Live Longer2
Why You Need to Ignore the Latest Misleading Study Glorifying Coffee3
You might think these headlines were from articles in different newspapers, but these dueling headlines are both from the New York Observer. Or, you might think that the articles were published months or years apart, but they appeared one day apart. Or, you might think that they're discussing different studies, but both are about the same two studies.
The article under the first headline is short and egregiously credulous, while the second is longer and appropriately skeptical. Since the second is dated the day after the first, it could almost serve as its retraction4. The Observer's new slogan could be: "Two newspapers for the price of one!"
Here's a sort of critical reader's digest version of both articles, side by side so that you can compare and contrast their treatments of the studies:
|Drinking Coffee Can Help You Live Longer||Why You Need to Ignore the Latest Misleading Study Glorifying Coffee|
|"Two new studies published in the Annals of Internal Medicine confirm that drinking coffee can lead to a longer life, so start sipping accordingly….||"…[I]s coffee healthy? … You may have heard of two large, recent studies that looked at exactly that question. And, like clock work, articles from across the world started screaming about coffee saving your life based on these very two studies. They were almost all wrong.|
|The first study, Coffee Drinking and Mortality in 10 European Countries: A Multinational Cohort Study, was quite comprehensive. The researchers looked at more than 520,000 people in 10 European countries and found that drinking more coffee could significantly lower a personís risk of mortality. … The second study, Association of Coffee Consumption With Total and Cause-Specific Mortality Among Nonwhite Populations, surveyed more than 185,000 African Americans, Native Americans, Hawaiians, Japanese Americans, Latinos and whites, and found that the above findings were still true. … Both studies separated the smokers and the nonsmokers, but found that it didnít change coffeeís effects. …||The two studies looked at a number of factors, one in a large U.S. population and another across a variety of European countries. … The researchers took truly massive samples of people5―one almost 200,000 and the other about 500,000―and compared their risks of death after taking into account a huge number of counfounding6 factors. They also looked at the effects of coffee across an enormous range of possible conditions, looking at everything from respiratory disease to suicide over a really impressive period of time. … After taking into account factors like smoking, drinking, weight, wealth, education and more, the scientists found that people who drank more coffee were less likely to die overall. They were also less likely to get any number of serious diseases…. Now, the researchers controlled for a lot of factors to try and iron out a definitive causative benefit from coffee. … But, to quote one of the study authors: "This is an observational study, We cannot say, OK, [if] you drink coffee it is going to prolong your life." … There are other issues with taking these studies as evidence that coffee is protective for health. As the authors note, there could be some reverse causality―people who are less healthy could self-select and choose to drink less coffee.|
|Regular coffee consumption helps ward off heart, respiratory and kidney diseases, in addition to strokes, diabetes and cancer. … Drinking two to four cups a day lowers your risk of death by 18 percent, in comparison to people who donít drink coffee at all."||These two studies were what is known as observational research. What this means is that the researchers took a group of people and artificially divided them up―in this case by the amount of coffee they drank―and compared these groups on bad outcomes like death. … There's a problem with observational research. You can't tell if the thing you're studying is causing the problems that you are seeing. It's possible that coffee makes you healthier, but it's also possible that this is just a correlation. Basically, there could be an underlying cause that makes people both drink more coffee and be healthy without coffee itself conferring a protective effect. …[H]onestly, the fact that it's observational research is enough to know that it's unlikely to prove that drinking coffee is going to save your life."|
- If you missed the first cup, see: Java Jive, 5/21/2012
- Margaret Abrams, "Drinking Coffee Can Help You Live Longer", Observer, 7/11/2017
- Gideon Meyerowitz-Katz, "Why You Need to Ignore the Latest Misleading Study Glorifying Coffee", Observer, 7/12/2017
- The second article doesn't explicitly mention the first article, but it does include a link to the first one when it discusses misreporting of the studies. So, clearly, this wasn't a case of the left hand not knowing what the right hand was doing; rather, the left hand knew what the right did, but didn't like it!
- This is misleading if not outright wrong, since it suggests that the cohorts studied were randomly selected just for these particular studies, which is incorrect. The first study used the EPIC―"European Prospective Investigation into Cancer and Nutrition"―cohort, which was recruited in the '90s. The second used the Multiethnic Cohort, which was established in the mid-'90s.
The idea of these large cohorts is to recruit a group of people, gather information about them, and then follow them over the course of many years, during which some will become ill or die. The researchers in these two studies simply accessed the large amounts of data gathered about these cohorts, and then used statistical analysis on them to compare the coffee drinkers to the non-drinkers. See:
- Sic: it should read "confounding".
Acknowledgment: The illustration is adapted from the cover of an old EC comic book.
The Puzzle of the Four Jacks
"The Four Jacks", as the police called them, was a gang of bank robbers, each of whom happened to have the first name "Jack". Their last names are, in alphabetical order: Clubb, Dimond, Hart, and Spayd*. However, according to police informant Eddie "the Snitch", there was a falling out between Jacks Clubb and Dimond that led to the break-up of the original gang.
"That's right," Eddie told Detective Davidson, "those two got into it and now they won't work together no more."
Now, another bank has been robbed, and eyewitnesses saw at least two masked robbers in the getaway car. According to Eddie, some of the Four Jacks are responsible for the crime, but he doesn't know who.
"Hart and Spayd are thick as thieves," Eddie informed Davidson, "they was cellmates in the pen, and you can bet that if either one of 'em was in on it the other was, too. But Spayd don't trust nobody, which is why he won't work with just one other guy, not even Hart. Clubb won't work alone with the two of 'em, 'cause he don't trust 'em. That's why he brought Dimond into the gang."
That's all that Eddie could tell Davidson. Assuming that what Eddie told him is correct, which if any of the four Jacks should Davidson arrest?
New Book: Skewed
I haven't read this new book yet, so I can't recommend it. However, I am intrigued by the book's subtitle: "A Critical Thinker's Guide to Media Bias". That's something I could use.
The topic of media bias is a difficult one to discuss in an unbiased way as it's so partisan: the right cries: "Liberal media bias!" and the left responds: "What liberal media?1 Fox! Rush!" However, both sides are right: there is liberal and conservative media bias, but it depends upon the media outlet. The New York Times and MSNBC2 are biased towards the left and the Democratic Party, while The Washington Times and Fox News Channel2 are biased to the right and the Republicans. Any minimally sophisticated consumer of the news knows this.
The author of the book, Larry Atkins, is a more-than-minimally sophisticated consumer of the news3; in fact, he's a professor of journalism. Unsurprisingly, he's also a liberal:
In an effort to be completely transparent, I should note that I am a staunch4 liberal. … In this book, however, I am attempting to give a balanced, objective, and centrist presentation…. I acknowledge that my attempt at balance may not be perfect, but I will try my very best to both acknowledge and critique conservative and liberal advocates.5
Fair enough. It would be unfair to demand perfection.
Like a previous New Book author6, he retracts the book's title in the "Introduction":
Don't be fooled by the title of this book. This will not be a diatribe against the so-called liberal media. … The point of this book is to examine media bias on both ends of the political spectrum, whether liberal or conservative.5
Well, that's good. I'm not too interested in reading a diatribe, though they can have their uses. One topic of the book appears to be what Atkins calls "advocacy journalism", which he describes as follows:
Advocacy journalism…is a type of reporting in which the reporter gives an opinion or point of view and uses stories to advance an agenda. … Yet many people rely on advocacy journalism as their main source of news. Perhaps most importantly, advocacy journalism has a polarizing effect upon society, and I hope this book deepens readers' understanding of this polarization. The book will also explore aspects of media literacy and ways to navigate the echo chamber of modern media. (By "echo chamber", I am referring to the tendency for people to get news and opinions from sources that "echo," or reinforce, their own or similar viewpoints.)5
Despite all that, Atkins adds: "I…wish to admit here and now that I am a fan of advocacy journalism, yet I know that it has its flaws, especially with respect to bias." So, I suppose that the book will be, at least in part, a defense of advocacy journalism. I look forward to reading this journalistic advocacy of advocacy journalism.
What worries me about advocacy journalism is whether the advocacy journalist is a journalist first and an advocate second, or an advocate first and a journalist second. If the former, then I have no problem with such journalism, even if I disagree with the position advocated―I'm not afraid of hearing people advocate positions I disagree with as long as they do so honestly. However, I wouldn't trust a journalist who put advocacy ahead of journalism―that's propaganda, not news.
In many cases, there is a conflict of interest between the advocate and the truth-seeker, so an "advocacy journalist" is in danger of having to choose between the role of advocate and that of journalist. If the truth goes against your cause, should you tell the truth or conceal it to protect the cause? Historically, there have been "advocacy journalists" who concealed the truth as well as those who exposed it even when it hurt their cause7.
There's also a paradox facing the advocacy journalist: an advocate wants to influence people in a certain way, but the ability to do so is partially dependent on the trust that comes from being an honest journalist. I'm not the only one who doesn't trust "journalists", such as Michael Moore, who are willing to put their advocacy ahead of the truth. So, what should the advocate who is also a journalist do? Pretend not to be an advocate? Or be an honest journalist, instead?
These are a few of the questions I hope to see addressed in Atkins' book.
- See Eric Alterman, What Liberal Media? The Truth About Bias and the News (2003)
- Among others. Also see note 4.
- "…[E]ven though I am liberal, I can recognize that news organizations such as MSNBC and various websites have a staunchly liberal agenda." ("Introduction") It's interesting that Atkins writes "even though" here, since it suggests that it wouldn't be surprising if he couldn't recognize it.
- Atkins likes the word "staunch": he uses it three times in a single paragraph, though once as an adverb (see previous note).
- See the "Introduction".
- See New Book: The Death of Expertise, 2/28/2017. Perhaps publishers ought to allow writers to entitle their own books.
- Compare and contrast the reporting of Walter Duranty and Malcolm Muggeridge on the Ukrainian famine. See Ron Radosh, "The mendacity of Walter Duranty", The New Criterion, 6/2012.
Solution to the Puzzle of the Four Jacks: Davidson should arrest Dimond, Hart, and Spayd.
There are sixteen subsets of the Four Jacks1, including the complete gang and the empty set2. We know from Eddie that the complete gang is no longer together, and we also know that someone from the gang is responsible for the robbery, so we can rule out the whole gang and the empty set. Four subsets of the gang consist of the four individual members of the original gang, but the eyewitness accounts indicate that there were at least two robbers, so we can rule out those four single-member subsets. That leaves ten possible subsets of the Four Jacks, using the obvious abbreviations:
- C, D, H
- C, D, S
- C, H, S
- D, H, S
- C, D
- C, H
- C, S
- D, H
- D, S
- H, S
1, 2, and 5 are ruled out, since Clubb and Dimond won't work together. 6, 7, 8, and 9 are ruled out because Hart and Spayd always work together. 3 is ruled out because Clubb won't work alone with Hart and Spayd. Finally, 10 is ruled out because Spayd won't work with just one other crook. That leaves, 4: the gang of Dimond, Hart, and Spayd robbed the bank.
- Any set of n members has 2n subsets, so that the Four Jacks has 24=16 subsets.
- Any set is a subset of itself, and the empty set is a subset of every set.