Table 3.

Major gun battles, community organizations, and migration. Logistic regression estimates of out-migration from Chitwan Valley

	Model 2 Market	Model 3 Farmer's coop	Model 4 Mill	Model 5 Health centre
Interactions
Gun battles*market	0.88* (2.19)
Gun battles*farmer's cooperative		0.73** (2.36)
Gun battles*mill			0.79*** (3.92)
Gun battles*health centre				0.89* (2.17)
Conflict
Gun battles (# per month)	1.16*** (3.50)	1.11*** (3.16)	1.20*** (4.66)	1.16*** (3.49)
Organization
Market	0.95 (0.83)	0.93 (0.98)	0.93 (1.01)	0.93 (1.01)
Farmer's cooperative	1.14 (1.19)	1.17^ (1.31)	1.14 (1.09)	1.14 (1.11)
Mill	1.02 (0.39)	1.02 (0.35)	1.05 (0.87)	1.02 (0.40)
Health centre	1.06^ (1.35)	1.06^ (1.37)	1.06^ (1.36)	1.08^ (1.63)
Police post	1.13 (0.87)	1.13 (0.88)	1.13 (0.85)	1.13 (0.86)
Residual	0.9597	0.9587	0.9584	0.9611
Neighbourhood level error variance	0.1070	0.1073	0.1070	0.1070
-2 res log likelihood	1,852,695	1,852,395	1,852,653	1,852,933

Notes:

1. Estimates are presented as odds ratios. Z statistics are given in parentheses.

2. No. of observations (person-months) = 258,333

3. All control variable used, but results not shown.

4. Community organizations are coded as accessible if they were within the following walking distance from the community: markets and mills- 5 minutes; farmer's cooperatives, health centres, and police posts- 10 minutes.

Statistical significance:

^{^}p<.10

^*p<.05

^**p<.01

^***p<.001 (one-tailed tests)