A basic human characteristic is our motivation to simplify the complex. It makes total sense that we like to reduce complexity as it takes less mental energy to consider a simple matter than a complex one. As discussed in a prior IFOD – we can only consider up to about seven points of data in our heads.
One way we simplify the complex is that we like single number ratings/rankings that encapsulate and simplify complex states. For example:
- We often think of our total cholesterol number as our cardiac risk even though there are many other factors at play such as HDL to LDL ratios, C-Reactive Protein levels, weight, family history, age,smoking habits, whether we eat meat, etc. that affect our cardiac risk.
- IQ tests purportedly boil down all the myriad aspects of intelligence into one number. As discussed in a prior IFOD, there are many different aspects to intelligence including visual-spatial, musical, interpersonal, intrapersonal, body-kinesthetic, linguistic, logical-mathematical, and naturalistic. A recent study found that personality had a bigger impact on career success than IQ.
- We look at Amazon ratings and prefer a book or product with a 4.8 star rating to a 4.2 rating. While a star rating does provide relevant information, the number of reviews and variation among reviews (both provided by Amazon) are needed to provide depth to those numbers. In order to really understand the ratings, however, it is necessary to read many of the reviews. Beyond reading the reviews, it would be even better to know which reviewers have tastes or preferences like you might have. For instance, some of my favorite books have a lot of five star ratings and one star ratings – so a “love it or hate it” book might be preferable in some cases than just across the board high ratings.
- Much like Amazon ratings, we like numerical scores for movies, music and other forms of entertainment. Why read a review when Rotten Tomatoes will tell you the movie is a 65 or Metacritic will report that an album is an 83?
- Credit reporting agencies produce a single FICO Score for borrowers based on payment history, current level of indebtedness, types of credit used, length of credit history, and new credit accounts. FICO scores have been heavily criticized for not capturing a borrower’s true ability or willingness to repay.
- We like numerical scores for the wine we buy (I guess I’ll spend a bit more on the 92 rated wine vs. the 90!). Of course, wine ratings are subjective and often flawed.
- We like numerical rankings of colleges. Universities spend millions (billions?) of dollars to move up a few spots in ranking because they know that it has a big effect on prospective students. Going to the number 9 ranked National University (Northwestern) seems a lot better than number 29 (NYU, Tufts, UNC, Georgia Tech, Rochester all tied at 29). Think about all the different aspects of a college experience and education including location, all the departments and degrees, sports, greek life, size, diversity of student population, cost etc. and it seems ludicrous that we give numerical ranks to colleges.
The point is not that simplification or looking to a single numerical summary score is bad, but rather we should be cognizant that a single number representing a complex topic like heart health, intelligence or whether The Goldfinch by Donna Tartt is a book you might like, is a very crude tool and digging deeper often makes sense.
A fascinating related topic: Goodhart’s Law