I read and liked Infinite Jest, yes. But you are missing the point, and you are making the error which is why I spent several paragraphs discussing what it means to 'control for' something by adding it as a covariate in a model. No one is surprised that IJ is often abandoned, either in absolute or proportional; but it is more surprising that the model is surprised by IJ's abandonment rates even after taking into account that it's by DFW and its current star-ratings and when it was published.
Star ratings and author don't really capture why a human observer would expect a lot of people to start and abandon this book though. You'd need some kind of measure of difficulty and/or trendiness.
