The Economics of Good Writing

Julio Cole and I have a new draft of our paper "Economical Writing (or, “Think Hemingway”)." We find that:

. . . Literature laureates tend to use shorter words than laureates in other disciplines, and the difference is statistically significant. These results confirm Salant’s idea that words are a scarce resource and should be used efficiently. This includes using short words instead of longer ones whenever possible. In short, good writing is also “economical writing.” 

Our main results: 

Table 4. Nobel Prize Banquet Speeches, 1969-2011 (summary statistics).
(a)  Syllables per word
	Econ	Physics	Chem	Med	Lit
Mean	1.456	1.440	1.470	1.472	1.344
Median	1.46	1.44	1.47	1.47	1.35
Maximum	1.63	1.61	1.68	1.63	1.68
Minimum	1.24	1.21	1.29	1.31	1.19
Std. Dev.	0.106	0.104	0.085	0.088	0.103
(b)  Characters per word
	Econ	Physics	Chem	Med	Lit
Mean	4.737	4.754	4.845	4.816	4.470
Median	4.75	4.77	4.80	4.82	4.49
Maximum	5.17	5.36	5.54	5.25	5.39
Minimum	3.97	4.05	4.39	4.20	4.09
Std. Dev.	0.297	0.288	0.248	0.250	0.254
Observations	42	45	42	47	27

Note that the lowest average word lengths are for the Literature prize. In terms of syllables/word the Literature laureates’ word lengths were, on average, almost 8% shorter than the weighted average for non-Literature laureates, and about 6.6% shorter in terms of characters/word. 

. . . 

ANOVA tests for only the non-Literature prizes (Table 6 [in the paper]) do not reject the hypothesis of equal means: average word length is essentially the same for Economics, Physics, Chemistry and Medicine laureates. Therefore, it is only the Literature laureates that can be regarded as a separate group when it comes to word length, which is consistent with the Salant hypothesis: the very best writers do indeed tend to use shorter words.

We conclude:

Salant (1969) argued that the use of short words is an indication of good writing. We found support for this hypothesis by comparing the banquet speeches of Nobel laureates. To be sure, word length is only one dimension of what makes for “good writing.” But it seems that it is a necessary dimension. Words are a scarce resource and must be used efficiently. This includes using short words rather than longer ones, whenever possible. “Economical” writing might indeed be the key for “good” writing. We leave for debate the different implications of our paper.