Tracking a factoid to its lair
« previous post | next post »
Matt Richtel, one of the leading current peddlers of the "technology is eating our brains" meme, is fond of this assertion:
The average person today consumes almost three times as much information as what the typical person consumed in 1960, according to research at the University of California, San Diego.
That version is the lead paragraph of the online site for his appearance on Fresh Air, "Digital Overload: Your Brain On Gadgets", 8/24/2010. I was curious about what this sentence could mean, and more specifically, I wondered which UCSD researchers did the measurements, and what they they measured. Usually I can track down the source of a factoid from the scant clues typically left by passing journalists, but this one has defeated me, so I'm asking for help.
The same claim is featured in Richtel's story "Attached to Technology and Paying a Price", NYT 6/6/2010:
For better or worse, the consumption of media, as varied as e-mail and TV, has exploded. In 2008, people consumed three times as much information each day as they did in 1960.
Later in the same article, Richtel writes that
At home, people consume 12 hours of media a day on average, when an hour spent with, say, the Internet and TV simultaneously counts as two hours. That compares with five hours in 1960, say researchers at the University of California, San Diego.
This may be the same factoid, in which case the measurement of quantity of information consumed in much cruder than I imagined, being denominated in hours of exposure (with or without attention actually being paid, apparently).
Anyhow, I think it would be a public service for someone to go through Richtel's NYT series and provide some footnotes linking his spectacular assertions about "research" to the publications behind them. (And isn't it past time for science and technology articles in reputable publications to provide such links as a matter of course?)
I don't know how this would all turn out in Richtel's case, but based on past experience, I have a suspicion.
[Update — a few additional Google Scholar probes turned up what is almost surely the source: Roger E. Bohn and James E. Short, "How Much Information? 2009 Report on American Consumers", Global Information Industry Center, UCSD. I'll have a bit more to say about this interesting document later on — meanwhile, feel free to discuss these issues in the comments, to whatever extent your poor information-riddled brains may permit.]
Faith said,
August 31, 2010 @ 10:43 am
I'm so enfeebled by information that I was under the impression that "3 times" five hours would be 15 hours, not 12. And the 12 hours is arrived at by counting some actual hours twice (if you have the computer and the tv on simultaneously), the reasoning for which also escapes me. I'd better go lie down with a cold laptop on my forehead.
Mark P said,
August 31, 2010 @ 10:50 am
What is the definition of information? According to my definition, when my eyes are open I am exposed to information. It might be more or less interesting than the information I get from watching television, but it is, nevertheless, information. The only time I am not exposed to information (or at least not consciously) is when I am asleep.
[(myl) Hey, don't be all vision-ist — what about acoustic and haptic information, huh?
Anyhow, FWIW, the cited HMI report (in Table 1 on p. 9) offers "Three Measures of Information":
As this display suggests, they're apparently only concerned with "information" presented in some kind of media-related package. As you note, this is absurd from either a psychological or an information-theoretic point of view, though clearly it's what matters if the only thing that you care about is consumption of media. In fact, of course, it's likely that someone who is taking a walk on the beach is actually taking in information at a faster rate than someone who is IMing a friend or checking their email.
Now that I think of it, you could make a case that the whole problem with modern media is too little sensory-motor information flow compared to the environment we evolved in, not too much. And on this view, the current trends are actually in a positive direction: in terms of rate of information flow, Grand Theft Auto is presumably more like life on the Pleistocene savannah than Pride and Prejudice is.]
Faldone said,
August 31, 2010 @ 10:53 am
@Faith
Your problem is that you're thinking in decimal. That's so last century. 3*5=F, which is 3 less than 12. So it's really four times the information.
Jac said,
August 31, 2010 @ 11:07 am
"And isn't it past time for science and technology articles in reputable publications to provide such links as a matter of course?"
I agree with this strongly. At least 3 times more strongly than I agreed with it in 1960.
MattF said,
August 31, 2010 @ 11:12 am
Well. Assuming 'consumes' = 'is exposed to', assuming no difference between signal and noise, assuming that response is a constant, universal, sensor-independent, and linear function of stimulus, assuming different information channels add together to produce a response in a linear, homogeneous, channel-independent fashion, then– I still don't believe it.
JFM said,
August 31, 2010 @ 11:25 am
From what I've had time to see/read, that report seems quite interesting.
Yes, they had a rather restricted definition of information. Still, the last sentence in their Executive Summary says: "Reading, which was in decline due to the growth of television, tripled from 1980 to 2008, because it is the overwhelmingly preferred way to receive words on the Internet." That's pretty close to the factoids in the Richtel quotes above, even though he says 1960, not 1980.
Anywho, the (claimed) fact that reading has tripled over three decades is a good thing, right? Taken at face value, it's positive, at least in terms of literacy skills/habits. (Although admittedly I haven't read the actual sections dealing with that yet, so perhaps there are some caveats there.)
Of course, it would be interesting also to see data on how that correlates with sociological factors (age, gender, education, profession, etc., etc.).
Jerry Friedman said,
August 31, 2010 @ 11:39 am
If you're watching commercial TV, you get information for, what, 40 minutes per hour? And even for fairly slow readers the pace of TV information is much slower than that of written information on the Internet, or if we're talking about 1960, books and newspapers. And most of it's not information anyway.
Spell Me Jeff said,
August 31, 2010 @ 11:50 am
An interesting sidestep from "information" to "media." I'm not sure these are interchangeable.
And isn't the media/non-media distinction a little artificial, anyway? We talk about the media as if it were a discrete thing. Is it? I get the distinction between artifact and not artificat. I'm just not sure it matters.
It's media if I watch it on Youtube. Most would agree.
It's not media if my gossipy neighbor tells me in person, even if the information is the same. Most would agree.
But what if my neighbor simply repeated something she learned on Youtube. Was my exposure a "mediated" exposure or not? If it isn't, what is the essential difference, and why should we care about it?
I go to a political rally and hear a politician speak. Not media, probably, even though it probably was amplified, and the text was produced on a computer. It's the nature of my personal exposure that counts, somehow.
I watch the same rally on TV. So I got it from the media. Despite the fact that a rally participant was do doubt exposed to a lot more information, in fact the kind of sensory/data overload that causes nausea and headaches. Again, why does the distinction matter?
More important is the value judgment implicit in the distinction. I'm not sure I'm ready to make such a judgment myself. Either way, the grounds for doing so seem shaky.
JFM said,
August 31, 2010 @ 11:54 am
I just feel the need to add that there's something fishy about the "word" count in the afore-mentioned report. I can't see any easily understandable explanation for how they transform consumed radio hours into number of consumed words.
[(myl) Well, you could take a representative sample of radio recordings, and count the words in transcripts, and extrapolate from that. I don't know whether that's what the HMI researchers did.]
AliL said,
August 31, 2010 @ 12:05 pm
If we take out all the bytes needed for downloading the ads that dominate so many webpages and appear every 5 minutes on TV, assuming this accounts for only 5% of the total space (which I consider a very low estimate), we have 180EB we really don't care about (which following their example on page 12, would make a pile of books 10cm high across the whole of the US+Alaska)
Secondly, taking video (TV and Computer Games) and saying that we can fullly interpret every pixel on a screen which FullHD has ~2M of them at ~30FPS is ridiculous.
Mary Kuhner said,
August 31, 2010 @ 12:16 pm
The industry definition of "information" in that article is way too far from any common-language use of the word to be quoted as such. They are measuring it in terms of bytes, and applying that definition across media, so their numbers (as they say themselves) are totally dominated by video and highly dependent on the resolution of the video. Apparently my son is absorbing massively more information playing Halo 3 than he was a few years ago playing its predecessor games, because there are more bytes in the video image. Of course, all video pales in comparison to, say, staring at the walls of his room at the full capacity of the human visual system (how many bytes is that?)
The conclusion from their approach would be that if I were playing solitaire on my college computer circa 1985 and had my ordinary TV on in the background, I would have been absorbing massively less information than if I were to play the exact same solitaire game with the exact same show on, but with a high-end modern graphics card and HDTV.
This definition is probably useful to someone, but to call it "information" is a stretch. I personally can't even tell the difference between HDTV and the regular kind. My visual acuity probably isn't up to it.
David Walker said,
August 31, 2010 @ 12:22 pm
In Table 1: How the heck does a human consume a "compressed byte"?
JFM said,
August 31, 2010 @ 12:24 pm
Ok, having read a little bit more, I see now that their Appendix B lists some numerical estimates, such as:
– Recorded music = 41 words per minute
– Radio = 80 words per minute
– Conventional telephone = 120 words per minute
– Newspapers = 240 words per minute
These seem to derive from other sources, and are accepted on faith/authority by the report's authors. I'm sure that's useful in order to make their study more easily comparable with other studies, but I would like to see what exactly those estimates are based on.
JimG said,
August 31, 2010 @ 1:20 pm
1)
> … person today consumes almost three times as much information
> as what the typical person consumed …
As WHAT? Am I the only reader who heard a loud clank?
2) Taking off from Jerry Friedman's idea, above, I'd say there's a big difference between information and noise, in any channel. From personal experience, the information content of "media" in the 1960s was much denser than in today's TV and print. With no definitive personal data, I might suggest an alternative hypothesis to that of the study and the interpretive article: Human brains, especially those belonging to science writers, have become mushier over the 50 years since 1960.
Josh said,
August 31, 2010 @ 1:30 pm
@David,
I expect that's their attempt to normalize for the density of information within a piece of media.
For example, imagine two 5MP photographs. One is a picture of a landscape with lots of detail and a high dynamic range (bright brights and dark darks). The other is a picture with the lens cap on. It's just black. In an uncompressed format, these two images are exactly the same size. However, if you compress them, the lens-cap picture will be much MUCH smaller. For at least video and audio media, this is probably a better approximation of "information consumed", whatever that ends up meaning.
I'm not so convinced that would work for text though. Written communication compresses REALLY well. If you took the same number of words from some Harry Potter novels as are in one of Dostoevsky's, they would compress to about the same size, but you'd have a hard time convincing me that the Harry Potter novels actually contained the same amount of information without some circular definition like information=number of words.
David Walker said,
August 31, 2010 @ 1:32 pm
Ugh… I hate the phrase "… as what the typical person consumed" as a shorthand for "the average of all persons" or something similar.
You could say that the typical family has 2.4 children, but NO families have that many children (I hope).
"Typical" and "the average of" are not the same thing, and I wish that that particular shorthand would just go away.
4ndyman said,
August 31, 2010 @ 2:26 pm
I usually have tunes pumping into my headphones while I'm busy on the computer at work. When I get home, I often spend still more time on the computer, but with the TV on, too.
It doesn't seem like a stretch by these measurement standards, then, that I could conceivably consume 24 hours of information a day, even though I get 6-7 hours of sleep a night.
Alex said,
August 31, 2010 @ 7:00 pm
Josh, I think your Harry Potter/Dostoevsky example is confusing prestige and aesthetics with information content. You wouldn't have to define information as word count to argue Rowling's work contains more "information." An argument: Rowling's work creates a society and laws of nature/magic out of whole cloth, while Dostoevsky merely described or left implicit those aspects in his work, because it is set in the real world.
I doubt the literary merits of information have much to do with our neurological capacity to process it.
groki said,
August 31, 2010 @ 7:59 pm
Spell Me Jeff:
I'm not sure which way the value judgment should go, but here's my take on the media/non-media distinction (and forgive me if this is obvious): it has to do with whether or not there's an editor of some sort involved, "mediating" your experience (or, if you like, standing in the way).
the creators of youtube videos and the producers of televised rallies restrict your experience to what they decide to post/broadcast. if you're physically there–experiencing it immediately–you can focus on whatever details you choose, not the details they find most salient or most photogenic or most "gotcha" or whatever.
now, your gossipy neighbor is also "mediating" your experience (and unless you live in a very talented neighborhood, your neighbor is not giving you the same experience watching the youtube yourself would). however, you can interact with your neighbor: ask questions, compare notes, gauge reliability of the information (and/or the neighbor).
the rally promoters are even more involved in mediating your experience: they try to flood you with their perspective (thus the nausea and headaches!). but if you're there you're free to notice whatever you want: you can, eg, watch the stage-side kerfuffle they won't broadcast, or gossip with your seatmates about the size of an entourage or the ugliness of a hairstyle, or complain about the chintziness of the snack bar.
in short: with media, someone who is not you decided what information you would be exposed to.
p.s. obviously, you can choose to watch other youtubes, change channels, google for contrasting views, etc. so, going "up a level," the web gives us unmediated (or at least semi-mediated) access to lots of media. yay web!
Harold said,
August 31, 2010 @ 9:27 pm
But, say, didn't Bill Kibben call this "the age of missing information"
http://www.billmckibben.com/age-of-missing-information.html
Mark Mandel said,
August 31, 2010 @ 9:48 pm
My wife Rene had brought the column to my attention, so I forwarded this post to her. Her reply:
The author's name is Matt Richtel. I would guess Mark Liberman conflated it with Frank Rich, who also writes for the NY Times, and/or his wife, Alex Witchel, who also writes for the Times.
[(myl) Oops. Careless of me to get his name wrong. Fixed now.]
Please note that he is reasonably respected in his profession (Times author profile page):
Matthew D. Richtel joined The New York Times in January 2000 as a technology reporter in the San Francisco bureau. Previously, Mr. Richtel was a freelance writer for various media outlets, including The Times. In 2010, he won the Pulitzer Prize for National Reporting for "Driven to Distraction," a series of articles on the troubling collision of 20th and 21st century technologies—driving and multitasking.
The series generated the biggest impact of anything The Times published in 2009. By the end of the year, state legislators had proposed more than 200 bills variously barring drivers from texting or phoning or requiring hands-free headsets. In an abrupt turnabout, cellphone companies and auto makers said they would support new curbs, and cellphone companies hurriedly mounted campaigns warning customers about the practice.
The term "distracted driving" became so familiar that Webster's New World chose it as its 2009 "Word of the Year".
[(myl) Driving (or operating other machinery) while texting is obviously a stupid thing to do, and it was clearly appropriate to warn about it with or without the support of scientific research. But in listening to Richtel's interview on Fresh Air, and reading through his Your Brain on Computers NYT articles, I was struck by the number of vague, unsourced and (to my mind) implausible assertions of allegedly scientific reasons to view modern life with alarm. These reminded me of the whole 2005 flap about "email lowers IQ more than pot", and of Nicholas Carr's equally unsupported campaign about the brain-rotting effects of hyperlinks.
So I started with the first thing in the Fresh Air piece, the business about how "The average person today consumes almost three times as much information as what the typical person consumed in 1960". What I've found so far is consistent with the general impression I took away from the interview, which is that Richtel is heaping up misunderstandings, exaggerations and inventions in order to make (or perhaps continue making) a splash about the socially and psychologically harmful effects of modern technology.]
Mark P said,
August 31, 2010 @ 11:06 pm
I didn't read very deeply into the Bohn and Short report, but I have one serious criticism. They talk about the amount of information consumed when they should talk about the amount of information delivered. I suspect that those are two very different quantities.
JFM said,
August 31, 2010 @ 11:28 pm
Mark P said,
>
>They talk about the amount of information consumed when
>they should talk about the amount of information delivered.
Even though they keep referring to "consumption", they are indded looking at delivered information. On p.10, for instance, they say: "we choose to define information as data that is delivered for use by a person."
Mark p said,
August 31, 2010 @ 11:53 pm
@JFM – Exactly.
Jerry Friedman said,
September 1, 2010 @ 12:55 am
@JFM: Thanks for correcting my false impression that they were only counting time, not trying to estimate bytes somehow.
Graeme said,
September 1, 2010 @ 6:45 am
Has the consumption of factoids soared accordingly?
Mark Mandel said,
September 1, 2010 @ 9:24 pm
Rene comments, inter alia:
Personally, I am amused by the idea that people can only deal with one complex issue at a time. Obviously this guy has never tried to cook dinner and clean up the kitchen at the same time. Not to mention write an essay while tending to a couple of preschoolers.
[(myl) Indeed. In a later post, I wrote:
]
maidhc said,
September 2, 2010 @ 5:10 am
If you want to consider consuming information in days past, I suggest reading Mark Twain's Life on the Mississippi and his description of the old riverboat pilots and how they maintained their knowledge of the river. It's a good read even if you're not particularly interested in this topic.
JimG said,
September 2, 2010 @ 11:55 am
I presume that there IS a limit to how much information a human brain can process, although I would assume that individual brains vary and that one might train to make more use of potential. Everything I've read, including this discussion and the cited items, leaves me confirmed in my view that there is no such thing as multitasking (defined as simultaneous processing and action). What we do, what computers do, is interruption and switching, combined with handing off tasks/actions to autonomous accessories. Thus, in cooking while parenting, we leave items on the heat, although we may mindlessly stir a sauce, while we talk. Texting while driving involves leaving the car to drive itself while we compose and input a message. The consequences of neglect may be omission of an ingredient, burned food, or death and distruction. The key concept seems to be how rapidly the brain can return to the former task, ideally before a problem can arise and become unrecoverable. Even when the multiple tasks are closely related as parts of a larger object or process, multitasking by humans can often lead to compromise, satisficing or error. We may be able to walk/drive and chew gum, but try carrying on two tasks involving language use of some complexity.
[(myl) There's a large and interesting psychological literature on cognitive load and resource competition, so that to some extent speculation in this area can be replaced by knowledge and experiment.]
ENKI-][ said,
September 2, 2010 @ 1:13 pm
This rather spurious factoid would be quite a bit more interesting from the point of view of information theory. It could be that the same amount of time was spent, but the media was more densely packed (or simply unpredictable, depending upon the model). I could certainly go for less predictable OR more densely packed media, but I really haven't seen much difference. Television and radio is just as formulaic and slow as it was in the sixties — maybe a little worse now that McLuhan's stuff is old hat ;-)
Jeff said,
September 2, 2010 @ 10:15 pm
I have subscribed to the NYT for 31 years. The one thing I have learned is that they should stop marking their sections "International", "National", "Business"… and create new sections called "Fiction" and "Non-Fiction". The Duke rape case is a showpiece. In their first article about the Lacrosse team party, they published physical evidence that the containing article could not possibly be true. I wrote them, pointing out that their first article was an obvious indication of a rogue prosecutor. I heard nothing, the NYT bathed in its story arc of "elite white vs. poor black", and now we find the prosecutor in jail. The NYT chooses its story arc, then ignores data. In this case, the NYT decided it wanted to do the "X-times the information" story arc, and then fingers hit the keyboard, facts not welcome.
Roger Bohn said,
September 3, 2010 @ 10:48 pm
This is a nice discussion of some of the paradoxes of trying to measure "information." In talking with numerous audience I find that this quickly turns into an emotional discussion, since we all have (multiple) definitions of "what is important." So in the end we just DEFINED three metrics, called INFOH, INFOW, and INFOC corresponding to hours, words, and (compressed) bytes. Everyone is free to come up with different definitions, and then repeat our study using their own measures!
To summarize, we attempted to count: artificially transmitted data reaching the eyeballs/eardrums. We did not try to adjust for "were they paying attention?" (As I facetiously put it, "If a TV plays in the next room with nobody watching, did a tree fall in the forest?") According to our accounting, it's certainly possible to have more than 24 hours in an information day – just keep the TV and radio on all day, then use a computer. We basically used Nielsen's definitions of hours consumed.
For INFOC I made a serious effort to calculate something close to a Shannon calculation of bits per second – see report for details. How many bits per second travel over the artificial communication channel to the media viewer (which could be a book, a TV set, etc.)? What many people would like for various purposes is some measure of VALUE of information. We could try to put $ value on some flows, but that's still a very restricted measure. Anything else requires very strong judgments – e.g. entertainment programs, versus news programs, and Fox News versus Daily Show versus advertisements. That's never going to be rigorous.
I agree that limiting our measurement to ARTIFICIAL information is not the only way to do things. I did a crude calculation that according to some metrics for "natural" information, some of our ancestors may have actually received more bytes than we do! But unfortunately I can't find historical data to make such calculations carefully, even if we overcame the conceptual issues.
As far as Matt Richtel's quotes about 1960 versus today, he asked us to generate some additional historical numbers that were not included in our report.
Rimero de enlaces said,
September 19, 2010 @ 3:56 am
[…] Tracking a factoid to its lair ¿Dónde nacen esos datos que todo el mundo repite como si fuesen cierto y realmente… bueno, no lo son? […]