Skip to main content
. 2012 Sep 5;7(9):e44068. doi: 10.1371/journal.pone.0044068

Table 3. Arithmetic and geometric means of water samples collected from adopters and non-adopters.

n Arithmetic mean (95% CI) Geometric mean (95% CI) p
Thermotolerant coliform (TTC) counts for household and source water
Household 490 153.2 (134.8; 171.6) 24.8 (19.9; 30.8) 0.75
Source 33 144.6 (78.8; 210.4) 29.9 (13.0; 68.7)
Source water quality (TTC) by adoption status *
Adopter 243 212.1 (188.4; 235.8) 65.3 (49.7; 85.8) 0.96
Non-adopter 247 213.4 (191.1; 235.7) 101.7 (82.4; 125.4)
Household water quality (TTC) by adoption status
Adopter 243 118.4 (95.4; 141.3) 13.7 (9.9; 18.8) 0.01
Non-adopter 247 187.5 (159.3; 215.7) 44.5 (33.7; 58.8)
Household water quality (TTC) by water treatment method
Confirmed Pureit use 186 86.6 (64.3; 108.8) 9.2 (6.5; 13.0) ref
Other filter methods 72 93.4 (53.9; 133.1) 15.5 (9.3; 25.9) 0.79
Reported boiling 24 209.7 (130.3; 289.1) 55.0 (19.2; 157.0) 0.04
Unsafe or no method 208 227 (194.7; 259.3) 64.4 (47.7; 86.8) <0.01
Log reduction in TTC from source to household by adoption status § *
Adopter 243 0.64 (0.47; 0.82) 0.24
Non-adopter 247 0.36 (0.21; 0.50)
Log reduction in TTC from source to household by water treatment method § *
Confirmed Pureit use 186 0.80 (0.61; 1.00) ref
Other filter methods 72 0.87 (0.58; 1.17) 0.59
Reported boiling 24 0.18 (−0.29; 0.64) 0.04
Unsafe or no method 208 0.13 (−0.01; 0.28) 0.01

p values determined by

§

T-tests or

negative binomial regression.

“ref” refers to the referent group.

*

Variance adjusted for clustering at source water sample.