Big Five personality traits and reading preferences

Book reading enthusiasts have an inclination to over-impute the inherent value of books and what they say about the reader. Or at least, they make claims that cannot currently be supported by empirical evidence. For example, one of the more frequent claims is that reading literary fiction makes you more empathetic to others. There is no evidence that this is true and there is a confounding variable - literary fiction tends to be quite a bit more popular among women than men and women tend to score higher on empathy tests than men. Consequently the causal flow might, and I suspect is, actually in the opposite direction. Reading literary fiction doesn't make you more empathetic, people who are more empathetic read literary fiction.

The other manifestation of the over-imputation of consequence to books are violence and pornography. It is assumed that if you read more violent books, it might induce greater violence in the person. In the post-World War II era this fear was initially centered on the damage violent comics might do to children. There were Senate hearings interrogating comic publishers (including William Gaines, later a founder of Mad Magazine.)

I have discussed the issue of the over-imputation of positive and negative affects from reading in many posts. The power of storytelling versus the power of the story is one which disaggregates the issue into constituent components. The belief in the power of books and book banning is another.

Setting aside this passionate belief in the power of books (with which I agree) to change lives in predictable ways (which I question), there is another persistent supposition which has rarely been rigorously measured. Do certain types of personalities prefer certain types of books? It sure sounds reasonable, but there are many reasonable things which are actually wrong.

Predicting Personality from Book Preferences with User-Generated Content Labels by Annalyn Ng, Maarten W. Bos, Leonid Sigal, and Boyang Li provides a clever approach towards testing this supposition. From the abstract:

Psychological studies have shown that personality traits are associated with book preferences. However, past
findings are based on questionnaires focusing on conventional book genres and are unrepresentative of niche content. For a
more comprehensive measure of book content, this study harnesses a massive archive of content labels, also known as ‘tags’, created by users of an online book catalogue, Goodreads.com. Combined with data on preferences and personality scores collected from Facebook users, the tag labels achieve high accuracy in personality prediction by psychological standards. We also group tags into broader genres, to check their validity against past findings. Our results are robust across both tag and genre levels of analyses, and consistent with existing literature. Moreover, user-generated tag labels reveal unexpected insights, such as cultural differences, book reading behaviors, and other non-content factors affecting preferences. To our knowledge, this is currently the largest study that explores the relationship between personality and book content references.

They acknowledge that there could be a self-selection issue for users of Facebook and GoodReads that might skew the data one way or another. This is somewhat offset by the unusually large data set. The results cannot be definitive but they are more strongly suggestive than anything else I have seen.

I think this is the most interesting chart among several.

Click to enlarge.

For each of the Big Five personality traits, there are genres of reading strongly associated with them. People who read a lot of chick lit, manifest strong extraversion whereas people who read a lot of manga tend to be introverted.

You can go down the list to see with which traits the types of books you read are associated and whether that matches your own self-assessment or what you have been told by others.

As an illustration, I read relatively little to almost none of the genres associated with Agreeableness and Extraversion. No manga, chick-lit, psychological drama, etc. And I would also judge that I am probably middle of the road on Agreeableness. On the other hand, I would have assessed myself as somewhat introverted even though I read no manga, very little-to-no supernatural, and very little horror, gothic, or parallel world. I do occasionally read some sci-fi and in the past I have read some fantasy but it is a small, small portion of what I read.

The other three personalities of the Big Five are Openness, Neuroticism and Conscientiousness. Here there is a much more predictive association for me. I read heavily among back burners (non-fiction that are of narrow interest to the general public), philosophical novels, university books, and classics. I don't read musicals but occasionally plays and certainly a lot of poetry (which I would suspect is in this category as well.) That pattern of reading puts me strongly in the Open aspect of openness and I suspect there would be a reasonably strong consensus among friends that that was correct.

I am light on theology-religion and business-leadership (I read them occasionally but not frequently) but am heavy on science and technology, politics and philosophy, and very heavy on professional reading. That puts me at the level-headed end of the Neuroticism scale. Again, I think there would be a strong consensus that that is a correct assessment.

I am heavy on four of the five categories that put one at the Conscientious end of the Conscientious scale: World War 2, professional reading, brain food, grown-up stuff. I am lighter on business and leadership but do read a reasonable amount in that category.

Using the scale created by Ng et al, based on what I read, I am highly Conscientious, very Open, very Levelheaded, and neither agreeable nor disagreeable. I think that all matches very well with how others might assess me. The only area where there is a marginal discrepancy is on Extraversion. The categories would indicate I am mildly introverted whereas I would self-assess as pretty solidly introverted.

A data point of one does not make a valid test of the tool, but it is an intriguing idea. If the matrix has merit, it would mean that librarians would be able to make genre recommendations based on measured personality. That would be pretty neat. I don't think the research is yet sufficiently robust to support that approach but it is adequate enough to pay attention to the idea.