Table 3.

Results of the classical panel data models for the log of TB incidence with meteorological factors of a 3-month lag

Factors	Simple linear regression			Fixed effects model			Random effects model
	Coefficient	t-value	P-value	Coefficient	t-value	P-value	Coefficient	t-value	P-value
Constant	1.0097	10.42	<0.0001				1.0367	12.27	<0.0001
MAT (°C)	-0.0195	-10.03	<0.0001	-0.0059	-4.96	<0.0001	-0.0060	-5.12	<0.0001
MP (mm)	0.0024	5.07	<0.0001	-0.0008	-2.77	0.0060	-0.0007	-2.63	0.0090
MSH (hours)	-0.0011	-2.56	0.0110	-0.0003	-1.10	0.2720	-0.0003	-1.14	0.2570
MAWS (m/s)	0.1441	5.94	<0.0001	0.0320	1.93	0.0550	0.0339	2.05	0.0410
Log likelihood	1.92			328.57
AIC	0.01			-1.32
SC	0.06			-1.21
F-statistic	193.90		<0.0001
H-statistic							10.41		0.0340
LM lag	14.14		<0.0001	39.40		<0.0001
Robust LM lag	161.22		<0.0001	21.00		<0.0001
LM error	34.13		<0.0001	35.93		<0.0001
Robust LM error	181.21		<0.0001	17.54		<0.0001
Moran’s I	0.20	6.07	<0.0001	0.20	6.20	<0.0001

TB tuberculosis, MAT monthly average temperature, MP monthly precipitation, MSH monthly total sunshine hours, MAWS monthly average wind speed, AIC Akaike information criterion, SC Schwarz Criterion, H-statistic Hausman-statistic, LM Lagrange multiplier