Table 6. Results of the negative binomial regression model for the reactions count.

	Parameter	B	Sig.	Exp(B)
	(Intercept)	-12.750	.000	0.001
Headline	Word length (avg.)	–0.206	.000	0.814***
	Word count (log2)	–0.008	.842	0.992
	Question count	–0.786	.398	0.456
	Unusual punctuation	0.550	.000	1.733***
	Common phrases	–0.259	.000	0.772***
	Unigrams	–0.077	.262	0.926
	Sentiment	0.173	.002	1.189**
Post text	Word length (avg.)	0.040	.034	1.041*
	Word count (log2)	0.063	.003	1.065***
	Question count	–0.239	.000	0.787***
	Unusual punctuation	–0.215	.027	0.807*
	Common phrases	0.057	.369	1.059
	Unigrams	0.334	.000	1.396***
	Sentiment	0.053	.300	1.055
Topics	Business	–0.006	.977	0.994
	Crime	0.007	.971	1.007
	Education	–0.344	.105	0.709
	Environment	–0.314	.248	0.730
	Travel	–0.672	.004	0.511**
	Media & Celebrities	0.164	.395	1.178
	Lifestyle	0.081	.675	1.084
	Politics	0.218	.267	1.244
	Sports	0.053	.812	1.054
Controls	Origin United States	0.087	.050	1.091*
	Tabloid	0.623	.000	1.864***
	Site follower (log2)	0.881	.000	2.414***
	(Scale)	1
	(Negative binomial)	1.638

* p ≤ .05

** p ≤ .01

*** p ≤ .001.