Table 3.

Correlations of daily bullish sentiment across sentiment measures

Panel A: Twitter
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)
Harvard-IV (1)	1.00	0.25	0.20	0.26	0.37	0.11	0.10	0.06	0.18
LM (2)		1.00	0.20	0.31	0.39	0.17	0.17	0.20	0.22
L1 (3)			1.00	0.50	0.25	0.30	0.31	0.05	0.38
L2 (4)				1.00	0.33	0.34	0.34	0.11	0.41
VADER (5)					1.00	0.24	0.21	0.19	0.33
Naive Bayes (6)						1.00	0.74	0.20	0.45
Max. entropy (7)							1.00	0.16	0.45
Deep-MLSA (8)								1.00	0.15
DeepMoji (9)									1.00

Panel B: StockTwits
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)
Harvard-IV (1)	1.00	0.28	0.19	0.32	0.41	0.20	0.24	0.15	0.23
LM (2)		1.00	0.22	0.31	0.37	0.16	0.18	0.23	0.21
L1 (3)			1.00	0.55	0.24	0.35	0.39	0.13	0.38
L2 (4)				1.00	0.36	0.48	0.52	0.18	0.53
VADER (5)					1.00	0.24	0.25	0.22	0.27
Naive Bayes (6)						1.00	0.76	0.14	0.57
Max. entropy (7)							1.00	0.14	0.57
Deep-MLSA (8)								1.00	0.15
DeepMoji (9)									1.00

Panel C: correlation between Twitter and StockTwits
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)
StockTwits-Twitter	0.26	0.32	0.27	0.33	0.31	0.22	0.22	0.25	0.29

This table reports correlations between daily bullish sentiment scores estimated from short messages published on Twitter and StockTwits based on different approaches (dictionary based and machine learning techniques). Panel A and B report correlations between daily bullish sentiment scores estimated from Twitter and StockTwits short messages, respectively. Panel C reports correlations between daily bullish sentiment scores estimated from Twitter short messages with those estimated from StockTwits short messages