. Author manuscript; available in PMC: 2018 Jun 1.

Published in final edited form as: J Water Sanit Hyg Dev. 2017 Jun 1;7(2):331–339. doi: 10.2166/washdev.2017.136

Table 1.

Results from two household-level regressions and one village-level regression to determine relation between probability of switching, education, distance to the nearest safe well, and distance to the HEALS area.

Independent variables of household-level regression	Probability of switching from unsafe wells	Independent variables of household-level regression	Probability of switching beyond nearest safe well	Independent variables of village-level regression	Probability of switching from unsafe wells
Education (year)	0.008 (p<0.001)	Education (yr)	0.008 (0.11)	Education (yr)	0.039 (0.00)
Distance to nearest safe well (m)	−0.001 (p<0.001)	Distance to nearest safe well (m)	−0.024 (p<0.001)	Distance to nearest safe well (m)	−0.001 (0.01)
				Log (distance to HEALS)	−0.058 (0.01)
				Constant	0.68 (0.00)
n	2003	n	467	n	76
Adjusted r²	0.22	Adjusted r²	0.11	Adjusted r²	0.28
Village effects included as controls for village differences in factors not observed		Village effects included as controls for village differences in factors not observed		F Statistic Prob > F	F(3, 72) = 10.7 0.0000