Jev's taste in books vs Goodreads readers

Does Jev like the books that Goodreads readers like, and where does it disagree most?

result2,980 questions

Jev's ranking of 2,980 books barely agrees with Goodreads readers' (rank correlation 0.19). It likes literary fiction and nonfiction such as Kazuo Ishiguro's Nocturnes and The Suspicions of Mr. Whicher far more than readers do, and scripture and fan-favorite genre fiction such as Dawn of Wonder far less.

How to read this: Each dot is a book. Across: its rank among Goodreads readers. Up: its rank for Jev. With agreement this low, the dots form a cloud rather than a line; the labeled ones are the biggest disagreements.

2,980 items; 90% interval 0.16 to 0.21.

In short

Jev and Goodreads readers order the same 2,980 books very differently, with a rank correlation of 0.19 where 1 would mean the identical order.
Jev ranks literary fiction and history above readers, and gives the Book of Mormon collection 0.97 of 4 where its readers average 3.36.
Goodreads fans rate books they chose to read, so part of the gap is who does the rating, not taste alone.

What the data shows

Put It Somewhere Else Patrick meme: the Goodreads top shelf; put it somewhere else, Patrick

Barely any agreement: rank correlation 0.19 (90% interval 0.16 to 0.21), the lowest of the five catalogs.
Jev likes serious, literary books more than readers do: Ishiguro's Nocturnes, The Suspicions of Mr. Whicher, Stacy Schiff's Cleopatra, Oedipus Rex and The Lost Child of Philomena Lee.
It likes what fans rate highest much less: the Book of Mormon collection (Jev 0.97 on the 0-to-4 scale, readers 3.36), the fantasy novel Dawn of Wonder (1.25 vs 3.36), and romance and series fiction such as The Golden Dynasty and Just One Drop.

What it means, and what it doesn't

Jev's book taste is a critic's: literary fiction, history, classics. Goodreads readers rate the books they love, and the ones they love most are often series, romance, fantasy and scripture that critics ignore. That's a real difference in what "good" means, and it's the biggest gap between Jev and any audience here.

It doesn't mean readers are wrong, or that Jev has actually read these books. The selection effect (fans rating their own favorites) explains part of the gap, and it would push any outside judge in the same direction.

Caveats

Fans rate what they chose to read. Goodreads users rate books they chose, often books they already expected to love. A devotional text or the fourth book in a romance series gets rated mostly by its fans; Jev rates everything cold. Part of the gap likely comes from that.
Stars squeezed into five levels. Goodreads uses 1 to 5 stars; each star is set on one of the project's five described levels to show them next to Jev's. The comparison is of ranks, not levels.
A popular-books list from 2017. The books are goodbooks-10k titles with at least 2,000 ratings, a popular, English-language list frozen in 2017.

Jev on this experiment

Would a person find it interesting to read?: Yes77%
Does it describe you?: No60%
Would you have predicted it?: No52%
How fair is the comparison?: The comparison is shaky
How much should a reader rely on it?: A little
Which caveat matters most?: Fans rate what they chose to read99%

Why ask this

Books have a huge, opinionated audience. Comparing Jev's ranking with real readers' ratings shows whether a model that has read about every one of these books likes the same ones the people who actually read them do.

It matters because "what should I read next?" is one of the most common things people ask a model. A recommender whose taste runs opposite to the readers of a genre will steer a fantasy or romance fan toward books they won't finish.

How this was done

The people and the data

The readers are Goodreads users, via goodbooks-10k, a public dataset collected in 2017 with millions of star ratings of the site's most-rated books, from bestsellers to scripture. The experiment keeps the 2,980 books with at least 2,000 ratings each, so every book here has a real crowd behind it. For each one it uses the whole spread of 1-to-5-star ratings, set on the same five levels Jev answers on.

What Jev was asked

Every book one at a time:

How much would you enjoy reading Shift by Hugh Howey?
You'd put it down after a chapter or two · You'd finish it out of duty and forget it soon after · You'd enjoy it once and not reread it · You'd recommend it to a friend · You'd reread it and count it among your favorite books

Each was also asked with the answers reversed, and the two averaged. Jev never saw the Goodreads ratings.

How it was measured

All 2,980 books are ranked by Jev's average level and by readers', and the orders are compared with a rank correlation (1 means the same order, 0 means none). Ranks, because Jev's described levels and people's stars aren't the same scale.

Where these questions live

2,980 questions across 1 topic of the map. Each opens on the map with every question in it.

Ratings

Book Ratings2,980

Every question

All 2,980 questions behind this result, the telling ones first: the examples the analysis points to, then the ones where Jev misses, biggest gap first.

Jev’s own answer

Showing 0 of 2,980