Reading habits of Americans

Here’s Pew’s report on the subject. I’ve included some numbers from the report below. Click to view tables in full size.

When interpreting the numbers in the post, do have in mind that people often lie about their reading habits. Also note that this may be an extremely biased sample; the cooperation rate is around 20% and the combined response rate was around 12,5% (1 in 8). I don’t know what kind of cooperation rate you usually get out of surveys like these, but my first instinct is to be critical seeing numbers like those; there may be a lot of stuff going on behind the scenes (out of sample). Anyway, ‘to the data…’:

About 1 in 5 didn’t read a book over the last 12 months, whereas about a third read 1-5. Approximately two-thirds read less than one book a month on average. But the interesting part to me was how many belonged in the category 11-50; about one in 4. That’s a lot more than I’d have thought. The 5% in the 50+ is also significantly higher than I’d have thought; I read 14 hours yesterday, I’m at maybe 30 hours this week so far (including all book reading, including ‘work book reading’ – though I must hasten to add that not all work is ‘book reading’) and I don’t think I’m even close to reading that many books over the course of a year. Of course ‘not all books are the same’, but even so. Not surprisingly the distribution is somewhat skewed: The mean was 17 books and the median was 8.

Females read more, old people read more and people who’ve been to college read more than people who have not. I thought it was interesting here that if you include education income is insignificant.

They do provide a more detailed picture of the number of books read. Of the people who read at least one book over the last 12 months, 8% had read one book, 17% had read 2-3 books, 16% had read 4-5 books, 19% had read 6-10, 18% had read 11-20 books and  22% had read more than 20 books. This means that 31% read 1-5 books per year (‘infrequent readers’), 28% read 6-20 (‘medium readers’) and only 17% (~1 in 6) of the population read more than 20 books per year (‘frequent readers’).

“A fifth of Americans (18%) said they had not read a book in the past year. This group is more likely to be: male than female (23% vs. 14%), Hispanic than white or black (28% vs. 17% and 16%), age 65 or older (27%), lacking a high school diploma (34%), living in households earning less than $30,000 (26%), unemployed (22%), and residents of rural areas 25%. Those who did not read a book last year also tended not to be technology users.”

I found none of those associations to be the least bit surprising.

58% regularly read daily news or newspapers. 48% regularly read magazines or journals. Even though it makes sense I did find it interesting that income becomes significant when looking at those numbers even when you account for education.

“Altogether, 43% of Americans age 16 and older have read long-form writing in digital format as of December 2011 – either e-books or newspaper or magazine material in digital form. We get that figure by combining those in the December survey who have read e-books with the 31% of those who regularly read news content and have read that content in digital format and the 16% who read magazines and journals and have read that content in digital format.
Those who have taken the plunge into reading e-books stand out in almost every way from other kinds of readers. They read more books than other readers. They read more frequently and are more likely than others to read for more purposes. They consume books in all formats, including print and audio: 88% of those who read e-books in the past 12 months also read print books. But they are also more likely than others to have bought their most recent book, rather than borrowed it, and they are more likely than others to say they prefer to purchase books.
Demographically, as of February 2012, the adults age 18 and older who read e-books are disproportionately likely to be under age 50, with higher levels of education and income.”

“64% of those 16 and older said they get book recommendations from family members, friends, or co-workers. […] 28% said they get recommendations from online bookstores or other websites. […] 23% said they get recommendations from staffers in bookstores they visit in person. […] 19% said they get recommendations from librarians or library websites.”

About one in four to one in five got recommendations from staffers in bookstores. I was surprised the number is that high. Some more detail: “Those most likely to get recommendations this way include: college graduates (28%), those living in households earning more than $75,000 (30%), parents of minor children (27%), technology owners and users, urban and suburban residents, and those under age 65.” The ‘college graduates’ really surprised me, as did the income one. The income one can perhaps be explained in terms of higher search costs because of higher opportunity cost of time but that’s not exactly convincing. Maybe it’s a ‘what are smart, rich, successful, educated people like me supposed to be reading these days’-effect?

I didn’t really care about the gadget stuff in the report but some of you may find that interesting.

September 27, 2012 - Posted by | Books, Data, Demographics


  1. Very good caveats. I’d like to add that people lie not only about their reading habits – most self-reported data is highly suspect, and much more so when it concerns behavior or characteristics that, if public, would enhance or damage a person’s status. In this, people often lie to themselves as well: “Hmmm, not sure how many books I read in the past year, but it’s just gotta be over 11 – after all, I’ve always liked books!”

    Now, putting on my grumpy number-cruncher hat on… Whiskey-tango-foxtrot is going on in the first table? First, who the hell designs a questionnaire in such a moronic fashion, with a grouping of 11-50? The first 2 non-“none” groups span 5 books/year, and the next one 40?! Big red flag right there – something fishy is going on in the 11-50 group. And indeed, between 1990 and 2011, the “none” category increased, all the others were flat or decreased, yet the mean went from 11 to 17 – a 50%+ increase! This, of course means there has to have been a tectonic shift within the 11-50 category. Technically, it’s possible, I guess, that the 1-5 groups average went from, say, 2 to 4, and the 6-10 groups average from, says, 7 to 9, to offset the “none” group’s increase, but seems unlikely, and even if so, the burden of the overall average transition has to be borne by the 11-50 group. It’s within-the-group average must have gone up wildly, and a breakdown into several more manageable groups would show it, but…. nah, why bother, right? To use their own breakdown, there must have been a massive outflow from the 11-20 books/year group and inflow into the 20+ group in those 20 years… but they do not show it, and let the reader deduce it, if he/she can.

    Like you, I also found their focus on e-books uninteresting, but I guess this is what the Bill & Melinda Gates Foundation paid for – fair enough.

    Comment by Plamus | September 28, 2012 | Reply

    • The table design may just be related to the table design of Gallup’s previous surveys. I wouldn’t think too much of that; they do give some breakdowns in the report so people are able to figure out at least some of the details if they care to. The jump in the number of books is one way to think of it; if you think in terms of percentages belonging to each group instead it’s maybe not an outrageous way to split up the data though it’s true that you miss out on a lot of the variation in the data. The 1-5 group is even larger than the 11-50 group.

      When it comes to the survey element and the self-reported data, I should maybe have been a bit more frank in the post; it seems obvious to me that there’s a high likelihood that some people are lying in the survey. 44-45% (depending on whether you include 16-17 year olds) said that they’d read a book the day before taking the survey (the PEW people interpret that as ‘read a book on an average day’). Yet more than two-thirds of the respondents read less than a book each month. Those numbers don’t add up. The report doesn’t even address this issue.

      Comment by US | September 28, 2012 | Reply

  2. “It’s within-the-group average” should, of course, read “Its within-the-group average” – my apologies.

    Comment by Plamus | September 28, 2012 | Reply

  3. “A fifth of Americans (18%) said they had not read a book in the past year. This group is more likely to be: male than female (23% vs. 14%), Hispanic than white or black (28% vs. 17% and 16%), age 65 or older (27%), lacking a high school diploma (34%), living in households earning less than $30,000 (26%), unemployed (22%), and residents of rural areas 25%. Those who did not read a book last year also tended not to be technology users.”

    You didn’t think the lack of a black-white difference is interesting? Makes me wonder if they corrected for intelligence. I’m having a hard time believing that whites who average some 15 points more do not read more. Does intelligence not predict reading habits?

    Also, these numbers are very high, but they counted both fiction and nonfiction, right? In that case, >50 is not that wild. I think I read 20-40 nonfiction books a year*. If spent that time reading fictions I could read hundreds. Fiction is just much faster.

    * This is a bad approximation of the total acquired information, cf.

    Comment by Emil | October 4, 2012 | Reply

    • Both blacks and whites have English as their native language and most of them have been through the American school system (rather than, say, the Mexican school system) – a lot of Hispanics living in the US don’t and have not. Given English as second language plus a significant proportion of them being migrant-workers working in mostly low-skilled, lower-than-average-income jobs I don’t find it surprising that more Hispanics than blacks don’t read at all.

      I’d expect blacks to read less than whites, and I think you can infer from the data that they do though I no longer remember if they calculated averages across race (income and education are important, and whites are better educated and earn more than blacks) – but I don’t see that as a strong argument for assuming that blacks should necessarily be significantly more likely to not read at all.

      “Makes me wonder if they corrected for intelligence.” – of course they didn’t, read the publication. How would you even do that given the sampling setup?

      Comment by US | October 4, 2012 | Reply

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: