What authors like versus what readers like

This week, under quarantine and working at home, I was pondering how any given author’s most popular work is popular for reasons other than what the author was interested in writing it for. The thread provided lots of excellent examples, like Upton Sinclair’s The Jungle [Amazon link], which is notorious for turning people into vegetarians instead of Communists.

Recently, I’ve also been interested in what what makes a good science fiction book good. There is a school of thought that it is the speculative aspect of science fiction that really defines it, which might be true in some sense, but I’m coming around to the idea that science fiction isn’t a genre at all, and is simply a subset of adventure fiction and romances, in the old use of the term.

So I’m going to put these two ideas together and look at some super popular twentieth century works of science fiction that have sequels or series behind them, and look to see whether readers found something different in the popular books than in their less popular sequels. My theory is that what authors are interested in is often quite different than what readers are interested in, and speculative science fiction is a really good test of this, since a popular book will encourage an author to write more of what he or she thinks made the book popular, but if readers find those ideas uninteresting or unappealing, the later books will be less popular.

I was curious if I could find any data on this. I decided to look at Goodreads star ratings for sequels to popular books, and as a control, I thought I would look at popular series in genres that have no speculative elements at all. I picked Agatha Christie mysteries and Patrick O’Brian’s Aubrey & Maturin series, two well-known sets of works that exist purely to give readers what they want.

Let’s start with Ender’s Game [Amazon link]. I was given Ender’s Game as an eighth-grader by a teacher who quite accurately surmised that I would like it. I went on to read the sequels, and was a little surprised by how different they were. Later, Card would regain some of the magic of Ender’s Game by retelling the story from the point-of-view of Bean [Amazon link], Ender’s foil in Battle School.

So let’s look at what Goodreads has to say about Ender’s Game, and the sequels, Speaker for the Dead, Xenocide, Children of the Mind, and then compare to Ender’s Shadow as a kind of internal control.

Ender’s Game gets 4.30 stars, Speaker for the Dead 4.07, Xenocide 3.79, Children of the Mind 3.76, a pretty clear downward trend with each. Let’s look at Ender’s Shadow: 4.31, almost exactly what Ender’s Game gets. Pretty good, for a story that is a retelling of another book he already wrote.

I’ve got a theory about why this is in Card’s case. Ender’s Game is an elaborate thinkpiece about genocide and military ethics, but I think most readers, myself included, don’t notice that at first, and find it an amazing bit of military sci-fi. As an interesting data point, consider that Jerry Pournelle put the short story version of “Ender’s Game” in the first volume his very long running There Will be War series. The later volumes get into the genocide theme explicitly, and work out speculative details of an interstellar human civilization, but just aren’t as fun.

Frank Herbert’s Dune is a book I really like, one of the most remarkable pieces of future history ever done, and hard sci-ish if you think of ecological concepts instead of spaceships. There was intrigue, interesting sociology, and the unforgettable Butlerian Jihad lurking in the background. And then later volumes had one of Paul Atreides descendants as a worm god. Shit got weird real quick.

I’m only going to consider books Frank wrote, as the shift to Frank’s son Brian will probably mask what I’m looking at.

Dune, 4.22 stars, Dune Messiah 3.88, Children of Dune 3.93, God Emperor of Dune [the worm one] 3.85, Heretics of Dune 3.86, Chapterhouse: Dune 3.91. Somewhat better than Card, Herbert at least stayed around the same level with his sequels. I think the space opera elements of Dune really carried it.

Larry Niven’s Ringworld kind of defines what I mean by hard sci-fi, but is also a pretty good example of how poorly speculative fiction ages. Larry’s future looks a lot like the 1970s when you look at it fifty years later. Again, I’m only going to look at books Larry wrote by himself.

Ringworld 3.96 stars, The Ringworld Engineers 3.87, The Ringworld Throne 3.54, Ringworld’s Children 3.75. Larry didn’t go down quite as much with volume 2, but volume 3 was a big drop in rating. I found The Ringworld Engineers pretty similar to Ringworld, so I think the ratings being pretty close makes sense.

I’m now going to include Jerry Pournelle’s Falkenberg series, as a military sci-fi control that seems more consistent over time. Let’s see what the data says. Again, I’m going to exclude the volumes written with S. M. Stirling, as I think that inserts variability that I’m not looking for.

The pattern here is different. The Mercenary 3.98 stars, West of Honor 4.04, Falkenberg’s Legion 4.15, Prince of Mercenaries 4.11. Jerry’s books go up as the series goes on, instead of down. Also, interestingly, the omnibus of all the books, including the ones co-authored with Stirling, gets a higher rating than any individual book. Possibly biased towards fans?

Now let’s turn to the outside controls, Christie and O’Brian.

Agatha Christie bounces around more than I expected. I am using Goodreads listing for sequence, as I am not a Christie fan. The Mysterious Affair at Styles 3.99 stars, The Murder on the Links 3.84, The Lemesurier Inheritance 3.87, Poirot Investigates 4.03, The Murder of Roger Ackroyd 4.24, The Big Four 3.59, The Mystery of the Blue Train 3.80, Black Coffee 3.58, Peril at End House 3.94, Lord Edgware Dies 3.91, Murder on the Orient Express 4.17. The list goes on, but you can see there is more variability here, star ratings both lower and higher than the other authors I looked at.

Patrick O’Brian’s Aubrey & Maturin series is more like Falkenberg’s Legion. The first book has the lowest rating, and things get better from there. Master and Commander 4.10 stars, Post Captain 4.29, H.M.S Surprise 4.42, The Mauritius Command 4.35, Desolation Island 4.40, The Fortune of War 4.40, The Surgeon’s Mate 4.39, The Ionian Mission 4.32, Treason’s Harbour 4.38, The Far Side of the World 4.45.

From this admittedly ad hoc survey, I did find some evidence in favor of what I was proposing. Lots of super famous speculative fiction novels do better than the sequels do, in Goodreads star rating terms. At least one series I thought was consistent over time actually was, Falkenberg’s Legion. Christie’s rating variability surprised me, but I am also not a fan. Maybe regular readers of hers wouldn’t be surprised. This counts against my theory. Patrick O’Brian blows everyone else of the water in star ratings.