(This post is mostly for teachers and librarians.)
I'm a big fan of Nate Silver's political analysis at
fivethirtyeight, a numbers-based blog that predicted the election results with astonishing accuracy.
I'm no Nate Silver, but I am interested in how writers, booksellers, librarians, and teachers can better understand the immense amount of information available to us online about how teenagers read, what they read, and (perhaps most importantly) what they like.
Our current methods of analysis almost always involve extremely small sample sizes--i.e., the teenagers at my library or school do/do not like this book; this book does/does not circulate in my library system; the teens on the BBYA panel thought this was derivative, or exciting, or deep, or funny, or lame, or too long.
This is of course very useful information, but it's ultimately anecdotal. We're asking ten or twenty teens to speak for ten or twenty thousand. The sample size is tiny, and too often we make broad conclusions from it.
So while
School Library Journal courts controversy by saying that children didn't like several recent Newbery winners, they don't feel compelled to provide any actual evidence to back up the claim.
Also, to use an example I've used many times before, adults often say that "their teens" do not like or get
Octavian Nothing. The conventional wisdom then becomes that
Octavian is a great book; i.e., that no one would ever choose to read it. But why then have
more than 100,000 people purchased it--making it a bigger commercial success than most "commercial" fiction published for teens?
I think we may have a way (albeit a still-flawed way*) of better quantifying reader appeal, and the results--to me, at least--were kind of stunning.
Okay. So there is hugely popular web site called
goodreads.** Most popular YA books have THOUSANDS of ratings on goodreads--so instead of dealing with ten kids, or a hundred kids, we're dealing with, say, three thousand kids. Also, they aren't mediating their ratings through us--aren't trying to meet or subvert our adult expectations. Here are some of the surprising and not surprising things one learns from goodreads ratings:
1. My least-liked novel is
An Abundance of Katherines--out of 2100 ratings, the average if 4.09 stars out of 5. (
Paper Towns is the best-liked of the books, with an average rating of 4.38;
Alaska is rated 4.36.)
2. If we assume that a rating of 4 or 5 stars constitutes an enjoyable read, a higher percentage of people report having enjoyed
Octavian Nothing 2 than report having enjoyed
Twilight. (Most of the
Octavian responders are teenagers, at least per my random checks.)
3. This year's (excellent!) Printz Award winner,
Jellicoe Road, has been called by some inaccessible to teen readers. The goodreads evidence indicates otherwise: With an admittedly-small-sample size of 122 ratings, its average rating of 4.43 is actually higher than any of the Printz honor books.
4. The most popular YA book of the year--ratings-wise--is Suzanne Collins's
The Hunger Games--which you may recall I told the New York Times
I liked a lot--with an average of 4.69 out of 3824 ratings.
5. Some books we consider commercially beloved may not be:
Looking for Alaska and
Gossip Girl 1 have almost the exact same number of ratings--50% of readers gave Gossip Girl 4 or 5 stars; 80% of readers gave 4 or 5 stars to
Alaska. (The same holds true for most of the books listed below; they just have fewer ratings.
(Of course all this begs the question: Should a writer's chief goal be an enjoyable reading experience?
Ulysses, for instance, has a lower average rating than any of my books; does this somehow mean that I'm a better writer than James Joyce? I think not, but that's for another post.)
I've listed some others below in case you're curious. I don't think metadata can replace the anecdotal reports librarians and teachers get from their readers, but I do think it can supplement those reports--and remind us all that the readers we have are not the only readers we
might have.
Paper Towns
1357 ratings
4.38 average
42% 5 stars
38% 4 stars
14% 3 stars
3% 2 stars
0% 1 star
An Abundance of Katherines
2105 ratings
4.09 average
31% 5 stars
41% 4 stars
20% 3 stars
4% 2 stars
1% 1 star
Looking for Alaska
3176 ratings
4.36 average rating
47% 5 stars
33% 4 stars
14% 3 stars
3% 2 stars
1% 1 star
Disreputable History of Frankie Landau-Banks
931 ratings
4.12 average
33% 5 stars
41% 4 stars
19% 3 stars
3% 2 stars
1% 1 star
Octavian Nothing 2
183 ratings
4.46 average rating
47% 5 stars
34% 4 stars
12% 3 stars
6% 2 stars
0% 1 star
Nation
563 ratings
4.30 average
46% 5 stars
39% 4 stars
11% 3 stars
1% 2 stars
0% 1 star
Twilight
105457 ratings
4.24 average rating
53% 5 stars
24% 4 stars
13% 3 stars
5% 2 stars
2% 1 star
Breaking Dawn
53415 ratings
4.05 average rating
45% 5 stars
26% 4 stars
16% 3 stars
7% 2 stars
4% 1 star
Jellicoe Road
122 ratings
4.43 average
57% 5 stars
26% 4 stars
8% 3 stars
3% 2 stars
4% 1 star
Hunger Games
3824 ratings
4.69 average rating
63% 5 stars
30% 4 stars
5% 3 stars
0% 2 stars
0% 1 star
Perks of Being a Wallflower
12473 ratings
4.03 average rating
36% 5 stars
34% 4 stars
20% 3 stars
6% 2 stars
2% 1 star
(Why don't these numbers add up to 100, you ask. It seems that goodreads always rounds down, meaning that the numbers can add up to as little as 96. Annoying, I know.)
* Said flaws begin with the fact that a lot of the people who rate YA books on goodreads are not, in fact, YAs. However, this may skew things in not-the-way-you'd-expect. At least from my cursory glance through reviews of
Paper Towns and
The Disreputable History..., teens seem to rate the books more highly than adults do.
But there are many other reasons not to just take goodreads ratings at face value. For instance, as a book gets more ratings, its average rating tends to sink, if only because a one-star review pulls things down so much. Above you'll see that the average rating for the contemporary classic
The Perks of Being a Wallflower are outrageously low. (
Speak is also low considering how much teens love it.) At first, I thought you could get around this by not counting the one-star ratings, but you actually can't, which makes me wonder whether maybe as a book gets out to a very wide audience, you just find more people who don't have very many enjoyable reading experiences.
Also, goodreads ratings obviously overvalue kids with high-speed internet access.
But none of this really distracts from what I'm trying to argue, which is that a LOT of teens REALLY LIKE the books that many people presume teens won't like, and that a lot of young people have more enjoyable reading experiences with ambitious work than they do with books that pander to their dreams and desires.
** I find goodreads' user interface annoying and prefer
librarything, but goodreads is much more popular with teens, who so far as I can tell enjoy annoying user interfaces.
