1 Introductory Notes
2 Dependent Variable: TTC
3 Independent Variables: Descriptive statistics and TTC correlations
4 Building a regression model – controls only
5 Building a regression model – selected IVs and controls
6 Building a regression model – All variables together

1 Introductory Notes

This file contains descriptive statistics and basic regressions for Byron’s H1, that more actively managed water sources will be less contaminated than less actively managed water sources. It also includes some attempts at full regression models.

2 Dependent Variable: TTC

This section contains descriptive statistics related to the TTC measure in the data. Here are summary statistics for TTC:

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max.    NA's 
##    0.00    0.00   10.25   41.29  107.25  125.00      20

NA or NA's refers to missing values.

Below is a histogram of the TTC variable, which ranges from 0 to 125, with 0 meaning no contamination and 125 meaning 125 or higher level of contamination.

As you can see, the vast majority of the readings are either 0 or 125 and over.

If we break down the TTC variable into categories of any contamination (TTC > 0) or no contamination (TTC = 0), we have the following distribution:

1=TTC level over 0, 0= TTC level is 0
Any Contamination?	Frequency		Total
Any Contamination?	Count	NA	Total
0	107 30.2 %	0 0 %	107 30.2 %
1	227 64.1 %	0 0 %	227 64.1 %
NA	20 5.6 %	0 0 %	20 5.6 %
Total	354 100 %	0 100 %	354 100 %

107 water sources have no contamination. 227 have contamination that is >0. The variable above will also be referred to as “TTC binary” in the rest of this file.

Below is TTC broken up into 10-point ranges:

##            
##             Count <NA>
##   (-Inf,50]   230    0
##   (50,100]     20    0
##   (100,150]    84    0
##   <NA>         20    0

##       
##        (-Inf,50] (50,100] (100,150] <NA>
##   0          107        0         0    0
##   0.5         10        0         0    0
##   1            3        0         0    0
##   1.5          4        0         0    0
##   2            5        0         0    0
##   2.5          2        0         0    0
##   3            3        0         0    0
##   3.5          1        0         0    0
##   4            1        0         0    0
##   4.5          3        0         0    0
##   5.5          3        0         0    0
##   6            2        0         0    0
##   6.5          3        0         0    0
##   7            4        0         0    0
##   7.5          6        0         0    0
##   8            2        0         0    0
##   9            2        0         0    0
##   9.5          5        0         0    0
##   10           1        0         0    0
##   10.5         2        0         0    0
##   11           2        0         0    0
##   11.5         2        0         0    0
##   12.5         1        0         0    0
##   13           2        0         0    0
##   14           1        0         0    0
##   14.5         1        0         0    0
##   15           3        0         0    0
##   15.5         3        0         0    0
##   16.5         1        0         0    0
##   17           2        0         0    0
##   17.5         1        0         0    0
##   18           1        0         0    0
##   19.5         1        0         0    0
##   20           2        0         0    0
##   21           2        0         0    0
##   21.5         1        0         0    0
##   22.5         2        0         0    0
##   24           2        0         0    0
##   24.5         1        0         0    0
##   25           3        0         0    0
##   25.5         4        0         0    0
##   26           1        0         0    0
##   29           1        0         0    0
##   29.5         1        0         0    0
##   30           2        0         0    0
##   34.5         2        0         0    0
##   37           1        0         0    0
##   37.5         1        0         0    0
##   38           1        0         0    0
##   40           1        0         0    0
##   40.5         2        0         0    0
##   42           2        0         0    0
##   44           2        0         0    0
##   45           1        0         0    0
##   46           1        0         0    0
##   47           1        0         0    0
##   50           3        0         0    0
##   51           0        2         0    0
##   53           0        1         0    0
##   54           0        2         0    0
##   56           0        2         0    0
##   57.5         0        1         0    0
##   60           0        1         0    0
##   68           0        1         0    0
##   69           0        1         0    0
##   74.5         0        1         0    0
##   81           0        1         0    0
##   84.5         0        1         0    0
##   87           0        1         0    0
##   87.5         0        1         0    0
##   89.5         0        1         0    0
##   91           0        1         0    0
##   91.5         0        1         0    0
##   99           0        1         0    0
##   110          0        0         1    0
##   113          0        0         1    0
##   121          0        0         1    0
##   125          0        0        81    0
##   <NA>         0        0         0   20

Grouped TTC
Contamination	Frequency		Total
Contamination	Count	NA	Total
(-Inf,50]	230 65 %	0 0 %	230 65 %
(50,100]	20 5.6 %	0 0 %	20 5.6 %
(100,150]	84 23.7 %	0 0 %	84 23.7 %
NA	20 5.6 %	0 0 %	20 5.6 %
Total	354 100 %	0 100 %	354 100 %

3 Independent Variables: Descriptive statistics and TTC correlations

This section contains a) tables or charts that describe some of our key independent variables and b) basic regressions of these variables against TTC readings.

3.1 Water Source Management Type (1.2.0.)

##                     
##                      Count <NA>
##   Borehole handpump    144    0
##   Borehole mech pump    22    0
##   Dug well open         46    0
##   Dug well handpump     71    0
##   Spring protected      40    0
##   Spring unprotected    10    0
##   Rain harvest tank     19    0
##   <NA>                   2    0

TTC vs Water Source Type, OLS
	TTC Round
Predictors	Estimates	CI	p
(Intercept)	30.07	21.94 – 38.19	<0.001
Borehole mech pump	2.93	-19.82 – 25.69	0.801
Dug well open	45.59	29.12 – 62.07	<0.001
Dug well handpump	15.16	0.92 – 29.41	0.038
Spring protected	7.85	-9.76 – 25.47	0.383
Spring unprotected	74.83	43.69 – 105.97	<0.001
Rain harvest tank	-16.29	-40.12 – 7.55	0.181
Observations	332
R² / adjusted R²	0.142 / 0.126

TTC vs Water Source Type, NBR
	TTC Round
Predictors	Incidence Rate Ratios	CI	p
(Intercept)	30.07	21.52 – 42.00	<0.001
Borehole mech pump	1.10	0.43 – 2.80	0.845
Dug well open	2.52	1.28 – 4.95	0.008
Dug well handpump	1.50	0.84 – 2.70	0.172
Spring protected	1.26	0.61 – 2.60	0.530
Spring unprotected	3.49	0.97 – 12.52	0.055
Rain harvest tank	0.46	0.17 – 1.23	0.121
Observations	332
Cox & Snell’s R² / Nagelkerke’s R²	0.046 / 0.067

TTC vs Water Source Type, ZINBR
	TTC Round
Predictors	Incidence Rate Ratios	CI	p
(Intercept)	57.21	55.49 – 58.98	<0.001
Borehole mech pump	1.05	0.97 – 1.14	0.255
Dug well open	1.62	1.54 – 1.69	<0.001
Dug well handpump	1.04	1.00 – 1.09	0.076
Spring protected	0.85	0.80 – 0.90	<0.001
Spring unprotected	1.83	1.71 – 1.96	<0.001
Rain harvest tank	0.48	0.42 – 0.55	<0.001
Zero-Inflated Model
(Intercept)	0.90	0.65 – 1.26	0.550
Borehole mech pump	0.91	0.35 – 2.33	0.838
Dug well open	0.25	0.11 – 0.57	0.001
Dug well handpump	0.35	0.18 – 0.68	0.002
Spring protected	0.31	0.13 – 0.72	0.006
Spring unprotected	0.00	0.00 – Inf	0.990
Rain harvest tank	1.11	0.41 – 2.96	0.838

## Number of observations in regression above:  332

Below is a logistic regression with the “any contamination” variable as the dependent variable:

1 = TTC is greater than 0
0 = TTC is equal to 0

TTC binary vs Water Source Type, Logistic
	TTC 0
Predictors	Odds Ratios	CI	p
(Intercept)	1.25	0.89 – 1.75	0.201
Borehole mech pump	0.98	0.38 – 2.52	0.968
Dug well open	4.24	1.77 – 10.18	0.001
Dug well handpump	2.73	1.40 – 5.32	0.003
Spring protected	3.44	1.41 – 8.37	0.006
Spring unprotected	12562276.42	0.00 – Inf	0.983
Rain harvest tank	1.26	0.46 – 3.45	0.651
Observations	332
Cox & Snell’s R² / Nagelkerke’s R²	0.089 / 0.124

Reference category is Borehole Hand Pump for above regressions.

3.2 Collected fees for water source (4.6.b.)

Data (water sources in Gulu) has been recoded into a variable with three categories:

Water sources that get 50% or more of their monthly collected fees from monthly fees.
Water sources that get 50% or more of their monthly collected fees from 1L jerrycans.
All other water sources (reference group in any regressions).

Below is a logistic regression with the “any contamination” variable as the dependent variable:

1 = TTC is greater than 0
0 = TTC is equal to 0

3.3 Risk of Contamination Score (Riskofcontamination10)

This variable is a score from 0 to 10 based on a visual inspection of a water source and how likely it is judged to be contaminated.

We start with a histogram showing the distribution of the variable:

Logistic regression:

TTC binary vs Risk of Contamination Score, Logistic
	TTC 0
Predictors	Odds Ratios	CI	p
(Intercept)	1.48	0.96 – 2.28	0.073
Riskofcontamination 10	1.08	1.00 – 1.17	0.057
Observations	332
Cox & Snell’s R² / Nagelkerke’s R²	0.011 / 0.015

4 Building a regression model – controls only

4.1 OLS for of TTC against controls

	TTC
Predictors	Estimates	CI	p
(Intercept)	42.69	26.54 – 58.83	<0.001
Borehole mech pump	8.35	-13.37 – 30.06	0.452
Dug well open	33.01	14.65 – 51.38	<0.001
Dug well handpump	4.57	-9.33 – 18.48	0.520
Spring protected	-0.57	-18.72 – 17.59	0.951
Spring unprotected	43.12	11.46 – 74.78	0.008
Rain harvest tank	-17.51	-40.13 – 5.10	0.130
Elevated	-35.14	-63.34 – -6.93	0.015
Flat	-18.50	-31.88 – -5.12	0.007
Undeveloped	24.60	9.99 – 39.20	0.001
Riskofcontamination 10	-0.78	-2.95 – 1.40	0.485
Observations	326
R² / adjusted R²	0.253 / 0.229

4.2 Logistic regression for TTC0¹ against controls:

	TTC 0
Predictors	Odds Ratios	CI	p
(Intercept)	2.11	0.95 – 4.70	0.068
Borehole mech pump	1.10	0.41 – 2.93	0.846
Dug well open	3.92	1.40 – 10.95	0.009
Dug well handpump	2.16	1.07 – 4.36	0.032
Spring protected	3.33	1.24 – 8.99	0.017
Spring unprotected	4503304.58	0.00 – Inf	0.984
Rain harvest tank	1.21	0.42 – 3.45	0.723
Elevated	0.53	0.14 – 2.04	0.356
Flat	0.58	0.30 – 1.11	0.101
Undeveloped	2.90	1.25 – 6.74	0.013
Riskofcontamination 10	0.92	0.83 – 1.03	0.159
Observations	326
Cox & Snell’s R² / Nagelkerke’s R²	0.148 / 0.207

4.3 Zero-inflated negative-binomial regression for TTC² against controls:

	TTC Round
Predictors	Incidence Rate Ratios	CI	p
(Intercept)	63.29	59.88 – 66.90	<0.001
Borehole mech pump	1.20	1.10 – 1.30	<0.001
Dug well open	1.21	1.14 – 1.29	<0.001
Dug well handpump	0.90	0.86 – 0.95	<0.001
Spring protected	0.72	0.68 – 0.77	<0.001
Spring unprotected	1.22	1.13 – 1.32	<0.001
Rain harvest tank	0.45	0.39 – 0.51	<0.001
Elevated	0.29	0.25 – 0.35	<0.001
Flat	0.75	0.71 – 0.78	<0.001
Undeveloped	1.27	1.21 – 1.33	<0.001
Riskofcontamination 10	1.01	1.00 – 1.02	0.025
Zero-Inflated Model
(Intercept)	0.53	0.24 – 1.17	0.118
Borehole mech pump	0.79	0.30 – 2.12	0.644
Dug well open	0.27	0.10 – 0.73	0.010
Dug well handpump	0.45	0.23 – 0.91	0.026
Spring protected	0.32	0.12 – 0.83	0.020
Spring unprotected	0.00	0.00 – Inf	0.990
Rain harvest tank	1.18	0.42 – 3.35	0.751
Elevated	1.65	0.43 – 6.29	0.465
Flat	1.75	0.92 – 3.33	0.086
Undeveloped	0.29	0.12 – 0.66	0.003
Riskofcontamination 10	1.09	0.98 – 1.22	0.120

5 Building a regression model – selected IVs and controls

This section contains OLS, Logistic, and ZINBR regressions for selected IVs plus controls.

5.1 Inspector Visits as IV

## [1] "########## OLS ##########"

	TTC
Predictors	Estimates	CI	p
(Intercept)	44.68	26.88 – 62.48	<0.001
`_2_3_inspectorvisits`	-0.18	-0.95 – 0.59	0.648
Borehole mech pump	9.93	-13.74 – 33.60	0.412
Dug well open	34.84	14.94 – 54.73	0.001
Dug well handpump	5.26	-10.04 – 20.56	0.501
Spring protected	-0.22	-20.05 – 19.60	0.982
Spring unprotected	44.72	11.53 – 77.91	0.009
Elevated	-38.20	-68.68 – -7.72	0.015
Flat	-18.25	-33.20 – -3.30	0.017
Undeveloped	22.83	6.52 – 39.14	0.006
Riskofcontamination 10	-1.03	-3.45 – 1.38	0.403
Observations	283
R² / adjusted R²	0.237 / 0.209

## [1] "########## LOGISTIC ##########"

	TTC 0
Predictors	Odds Ratios	CI	p
(Intercept)	2.37	0.99 – 5.64	0.052
`_2_3_inspectorvisits`	0.95	0.80 – 1.13	0.569
Borehole mech pump	1.02	0.36 – 2.87	0.972
Dug well open	3.55	1.24 – 10.17	0.019
Dug well handpump	2.04	0.96 – 4.31	0.062
Spring protected	2.75	0.99 – 7.60	0.051
Spring unprotected	4848648.04	0.00 – Inf	0.984
Elevated	0.46	0.11 – 1.83	0.268
Flat	0.58	0.29 – 1.19	0.136
Undeveloped	2.37	0.98 – 5.73	0.055
Riskofcontamination 10	0.93	0.82 – 1.04	0.198
Observations	283
Cox & Snell’s R² / Nagelkerke’s R²	0.143 / 0.201

## [1] "########## ZINBR ##########"

	TTC Round
Predictors	Incidence Rate Ratios	CI	p
(Intercept)	63.86	60.17 – 67.77	<0.001
`_2_3_inspectorvisits`	0.99	0.97 – 1.01	0.214
Borehole mech pump	1.26	1.16 – 1.37	<0.001
Dug well open	1.31	1.23 – 1.39	<0.001
Dug well handpump	0.93	0.88 – 0.98	0.004
Spring protected	0.77	0.72 – 0.83	<0.001
Spring unprotected	1.30	1.20 – 1.41	<0.001
Elevated	0.27	0.22 – 0.33	<0.001
Flat	0.80	0.76 – 0.84	<0.001
Undeveloped	1.26	1.20 – 1.33	<0.001
Riskofcontamination 10	1.00	0.99 – 1.01	0.960
Zero-Inflated Model
(Intercept)	0.48	0.20 – 1.11	0.087
`_2_3_inspectorvisits`	1.04	0.91 – 1.19	0.590
Borehole mech pump	0.84	0.30 – 2.36	0.739
Dug well open	0.30	0.11 – 0.83	0.020
Dug well handpump	0.43	0.21 – 0.92	0.029
Spring protected	0.39	0.14 – 1.05	0.061
Spring unprotected	0.00	0.00 – Inf	0.990
Elevated	1.98	0.50 – 7.87	0.333
Flat	1.82	0.90 – 3.65	0.093
Undeveloped	0.38	0.16 – 0.90	0.028
Riskofcontamination 10	1.08	0.96 – 1.22	0.182

## Number of observations in ZINBR regression above:  283

5.2 Number of regular users as IV

## [1] "########## OLS ##########"

	TTC
Predictors	Estimates	CI	p
(Intercept)	42.76	24.06 – 61.45	<0.001
`_4_2_regularuserhhs`	0.00	-0.02 – 0.03	0.885
Borehole mech pump	8.90	-16.71 – 34.51	0.496
Dug well open	33.30	11.76 – 54.83	0.003
Dug well handpump	6.66	-9.27 – 22.60	0.413
Spring protected	2.99	-18.83 – 24.81	0.788
Spring unprotected	42.83	8.63 – 77.02	0.015
Elevated	-39.44	-75.02 – -3.85	0.031
Flat	-16.24	-31.73 – -0.75	0.041
Undeveloped	25.39	8.72 – 42.07	0.003
Riskofcontamination 10	-0.89	-3.53 – 1.74	0.506
Observations	261
R² / adjusted R²	0.224 / 0.193

## [1] "########## LOGISTIC ##########"

	TTC 0
Predictors	Odds Ratios	CI	p
(Intercept)	1.83	0.76 – 4.41	0.178
`_4_2_regularuserhhs`	1.00	1.00 – 1.00	0.255
Borehole mech pump	1.18	0.40 – 3.49	0.769
Dug well open	3.14	1.01 – 9.75	0.048
Dug well handpump	2.04	0.96 – 4.35	0.065
Spring protected	2.35	0.77 – 7.14	0.131
Spring unprotected	4562797.56	0.00 – Inf	0.984
Elevated	0.32	0.06 – 1.58	0.161
Flat	0.60	0.29 – 1.23	0.162
Undeveloped	2.20	0.91 – 5.34	0.081
Riskofcontamination 10	0.96	0.85 – 1.09	0.523
Observations	261
Cox & Snell’s R² / Nagelkerke’s R²	0.132 / 0.186

## [1] "########## ZINBR ##########"

	TTC Round
Predictors	Incidence Rate Ratios	CI	p
(Intercept)	63.54	59.92 – 67.37	<0.001
`_4_2_regularuserhhs`	1.00	1.00 – 1.00	0.007
Borehole mech pump	1.21	1.10 – 1.33	<0.001
Dug well open	1.25	1.17 – 1.33	<0.001
Dug well handpump	0.94	0.89 – 0.99	0.022
Spring protected	0.84	0.78 – 0.90	<0.001
Spring unprotected	1.25	1.15 – 1.35	<0.001
Elevated	0.31	0.24 – 0.38	<0.001
Flat	0.81	0.77 – 0.85	<0.001
Undeveloped	1.34	1.28 – 1.41	<0.001
Riskofcontamination 10	1.00	0.99 – 1.01	0.962
Zero-Inflated Model
(Intercept)	0.62	0.26 – 1.47	0.274
`_4_2_regularuserhhs`	1.00	1.00 – 1.00	0.284
Borehole mech pump	0.72	0.24 – 2.14	0.556
Dug well open	0.28	0.09 – 0.86	0.027
Dug well handpump	0.43	0.20 – 0.91	0.029
Spring protected	0.45	0.15 – 1.31	0.143
Spring unprotected	0.00	0.00 – Inf	0.990
Elevated	2.87	0.58 – 14.25	0.197
Flat	1.73	0.85 – 3.54	0.133
Undeveloped	0.41	0.17 – 0.99	0.046
Riskofcontamination 10	1.05	0.92 – 1.18	0.489

## Number of observations in ZINBR regression above:  261

Bivariate OLS of just TTC vs number of regular users, which noteworthily shows no relationship of any kind!

	TTC
Predictors	Estimates	CI	p
(Intercept)	45.25	37.82 – 52.68	<0.001
`_4_2_regularuserhhs`	0.01	-0.02 – 0.04	0.585
Observations	264
R² / adjusted R²	0.001 / -0.003

5.3 _4_3_a_authorizedlist as IV

## [1] "########## OLS ##########"

	TTC
Predictors	Estimates	CI	p
(Intercept)	45.35	27.80 – 62.89	<0.001
`_4_3_a_authorizedlist`	-7.21	-19.88 – 5.45	0.265
Borehole mech pump	1.91	-21.01 – 24.84	0.870
Dug well open	28.69	8.69 – 48.70	0.005
Dug well handpump	5.01	-9.64 – 19.66	0.503
Spring protected	-0.07	-19.29 – 19.14	0.994
Spring unprotected	47.17	12.82 – 81.52	0.008
Elevated	-35.23	-64.25 – -6.21	0.018
Flat	-18.95	-33.25 – -4.66	0.010
Undeveloped	24.68	9.20 – 40.17	0.002
Riskofcontamination 10	-0.62	-2.96 – 1.71	0.601
Observations	293
R² / adjusted R²	0.249 / 0.222

## [1] "########## LOGISTIC ##########"

	TTC 0
Predictors	Odds Ratios	CI	p
(Intercept)	2.16	0.92 – 5.05	0.077
`_4_3_a_authorizedlist`	0.77	0.42 – 1.41	0.400
Borehole mech pump	0.90	0.33 – 2.47	0.841
Dug well open	3.24	1.11 – 9.42	0.031
Dug well handpump	2.09	1.01 – 4.32	0.046
Spring protected	2.88	1.06 – 7.84	0.038
Spring unprotected	4343738.46	0.00 – Inf	0.985
Elevated	0.57	0.15 – 2.20	0.416
Flat	0.63	0.32 – 1.25	0.188
Undeveloped	2.70	1.14 – 6.38	0.023
Riskofcontamination 10	0.94	0.84 – 1.05	0.292
Observations	293
Cox & Snell’s R² / Nagelkerke’s R²	0.139 / 0.195

## [1] "########## ZINBR ##########"

	TTC Round
Predictors	Incidence Rate Ratios	CI	p
(Intercept)	66.83	63.07 – 70.82	<0.001
`_4_3_a_authorizedlist`	0.84	0.81 – 0.88	<0.001
Borehole mech pump	1.04	0.95 – 1.15	0.356
Dug well open	1.14	1.07 – 1.22	<0.001
Dug well handpump	0.90	0.86 – 0.95	<0.001
Spring protected	0.76	0.71 – 0.81	<0.001
Spring unprotected	1.25	1.15 – 1.36	<0.001
Elevated	0.29	0.24 – 0.34	<0.001
Flat	0.75	0.71 – 0.79	<0.001
Undeveloped	1.27	1.22 – 1.33	<0.001
Riskofcontamination 10	1.01	1.00 – 1.02	0.026
Zero-Inflated Model
(Intercept)	0.54	0.23 – 1.25	0.149
`_4_3_a_authorizedlist`	1.10	0.60 – 2.02	0.747
Borehole mech pump	0.93	0.34 – 2.55	0.888
Dug well open	0.30	0.11 – 0.86	0.025
Dug well handpump	0.42	0.20 – 0.87	0.020
Spring protected	0.36	0.14 – 0.95	0.038
Spring unprotected	0.00	0.00 – Inf	0.991
Elevated	1.59	0.42 – 6.08	0.497
Flat	1.66	0.85 – 3.25	0.140
Undeveloped	0.33	0.14 – 0.77	0.011
Riskofcontamination 10	1.07	0.96 – 1.20	0.235

## Number of observations in ZINBR regression above:  293

5.4 _4_4_feecollectionsystem as IV

This is pretty fascinating: _4_4_feecollectionsystem is not significant in OLS and logistic but is singificant in ZINBR. We should keep an eye on this in other logistic regressions. This may be our best example so far of the added value of doing ZINBR.

## [1] "########## OLS ##########"

	TTC
Predictors	Estimates	CI	p
(Intercept)	46.21	27.81 – 64.62	<0.001
`_4_4_feecollectionsystem`	-5.40	-17.49 – 6.68	0.382
Borehole mech pump	8.27	-14.85 – 31.39	0.484
Dug well open	28.33	7.58 – 49.07	0.008
Dug well handpump	4.04	-10.51 – 18.60	0.586
Spring protected	-5.20	-25.09 – 14.70	0.609
Spring unprotected	37.69	4.06 – 71.33	0.029
Elevated	-35.79	-64.77 – -6.82	0.016
Flat	-19.06	-33.27 – -4.86	0.009
Undeveloped	26.31	10.86 – 41.77	0.001
Riskofcontamination 10	-0.67	-3.00 – 1.65	0.571
Observations	297
R² / adjusted R²	0.249 / 0.223

## [1] "########## LOGISTIC ##########"

	TTC 0
Predictors	Odds Ratios	CI	p
(Intercept)	2.00	0.82 – 4.85	0.126
`_4_4_feecollectionsystem`	1.03	0.57 – 1.85	0.926
Borehole mech pump	1.19	0.43 – 3.35	0.736
Dug well open	3.84	1.25 – 11.74	0.018
Dug well handpump	2.01	0.98 – 4.12	0.055
Spring protected	3.04	1.07 – 8.62	0.036
Spring unprotected	4671230.24	0.00 – Inf	0.984
Elevated	0.55	0.14 – 2.13	0.390
Flat	0.62	0.32 – 1.22	0.169
Undeveloped	2.83	1.20 – 6.68	0.017
Riskofcontamination 10	0.93	0.83 – 1.04	0.212
Observations	297
Cox & Snell’s R² / Nagelkerke’s R²	0.140 / 0.197

## [1] "########## ZINBR ##########"

	TTC Round
Predictors	Incidence Rate Ratios	CI	p
(Intercept)	69.07	64.93 – 73.47	<0.001
`_4_4_feecollectionsystem`	0.86	0.83 – 0.90	<0.001
Borehole mech pump	1.14	1.04 – 1.24	0.004
Dug well open	1.10	1.03 – 1.17	0.005
Dug well handpump	0.90	0.86 – 0.95	<0.001
Spring protected	0.66	0.61 – 0.71	<0.001
Spring unprotected	1.10	1.01 – 1.19	0.029
Elevated	0.29	0.24 – 0.35	<0.001
Flat	0.75	0.71 – 0.78	<0.001
Undeveloped	1.31	1.25 – 1.37	<0.001
Riskofcontamination 10	1.01	1.00 – 1.02	0.027
Zero-Inflated Model
(Intercept)	0.58	0.24 – 1.38	0.217
`_4_4_feecollectionsystem`	0.90	0.50 – 1.62	0.725
Borehole mech pump	0.70	0.25 – 1.97	0.503
Dug well open	0.26	0.09 – 0.77	0.016
Dug well handpump	0.44	0.21 – 0.89	0.023
Spring protected	0.34	0.12 – 0.94	0.037
Spring unprotected	0.00	0.00 – Inf	0.990
Elevated	1.62	0.42 – 6.21	0.479
Flat	1.72	0.88 – 3.35	0.112
Undeveloped	0.32	0.14 – 0.75	0.009
Riskofcontamination 10	1.08	0.96 – 1.21	0.181

## Number of observations in ZINBR regression above:  297

5.5 _4_17_planOandM as IV

## [1] "########## OLS ##########"

	TTC
Predictors	Estimates	CI	p
(Intercept)	36.47	15.46 – 57.49	0.001
`_4_17_planOandM`	2.60	-9.46 – 14.66	0.673
Borehole mech pump	9.90	-13.15 – 32.94	0.401
Dug well open	37.02	16.98 – 57.05	<0.001
Dug well handpump	6.41	-8.77 – 21.58	0.409
Spring protected	2.03	-17.08 – 21.15	0.835
Spring unprotected	44.04	10.97 – 77.10	0.010
Elevated	-33.51	-64.99 – -2.03	0.038
Flat	-14.34	-29.13 – 0.45	0.058
Undeveloped	28.03	11.95 – 44.11	0.001
Riskofcontamination 10	-0.66	-3.06 – 1.74	0.590
Observations	287
R² / adjusted R²	0.248 / 0.221

## [1] "########## LOGISTIC ##########"

	TTC 0
Predictors	Odds Ratios	CI	p
(Intercept)	1.27	0.46 – 3.48	0.645
`_4_17_planOandM`	1.35	0.74 – 2.49	0.329
Borehole mech pump	1.25	0.45 – 3.45	0.665
Dug well open	4.20	1.39 – 12.74	0.011
Dug well handpump	2.15	1.03 – 4.52	0.042
Spring protected	3.14	1.16 – 8.54	0.025
Spring unprotected	4642130.50	0.00 – Inf	0.984
Elevated	0.58	0.13 – 2.49	0.462
Flat	0.68	0.34 – 1.38	0.288
Undeveloped	3.17	1.29 – 7.82	0.012
Riskofcontamination 10	0.96	0.85 – 1.08	0.486
Observations	287
Cox & Snell’s R² / Nagelkerke’s R²	0.151 / 0.212

## [1] "########## ZINBR ##########"

	TTC Round
Predictors	Incidence Rate Ratios	CI	p
(Intercept)	59.73	55.73 – 64.01	<0.001
`_4_17_planOandM`	1.02	0.98 – 1.06	0.304
Borehole mech pump	1.17	1.08 – 1.28	<0.001
Dug well open	1.33	1.24 – 1.41	<0.001
Dug well handpump	0.93	0.88 – 0.98	0.007
Spring protected	0.79	0.74 – 0.85	<0.001
Spring unprotected	1.29	1.19 – 1.40	<0.001
Elevated	0.31	0.26 – 0.38	<0.001
Flat	0.82	0.78 – 0.87	<0.001
Undeveloped	1.33	1.27 – 1.39	<0.001
Riskofcontamination 10	1.00	0.99 – 1.01	0.670
Zero-Inflated Model
(Intercept)	0.78	0.29 – 2.14	0.635
`_4_17_planOandM`	0.82	0.45 – 1.51	0.525
Borehole mech pump	0.69	0.25 – 1.89	0.465
Dug well open	0.27	0.09 – 0.78	0.015
Dug well handpump	0.42	0.20 – 0.88	0.022
Spring protected	0.34	0.13 – 0.90	0.030
Spring unprotected	0.00	0.00 – Inf	0.990
Elevated	1.55	0.36 – 6.65	0.555
Flat	1.57	0.79 – 3.13	0.198
Undeveloped	0.29	0.12 – 0.70	0.006
Riskofcontamination 10	1.05	0.93 – 1.18	0.415

## Number of observations in ZINBR regression above:  287

5.6 _4_18_enoughfundsrepair as IV

## [1] "########## OLS ##########"

	TTC
Predictors	Estimates	CI	p
(Intercept)	41.47	22.94 – 59.99	<0.001
`_4_18_enoughfundsrepair`	-4.09	-16.66 – 8.47	0.524
Borehole mech pump	10.83	-12.09 – 33.76	0.355
Dug well open	31.84	11.94 – 51.73	0.002
Dug well handpump	2.55	-12.47 – 17.57	0.740
Spring protected	0.90	-18.55 – 20.34	0.928
Spring unprotected	42.09	9.21 – 74.96	0.013
Elevated	-30.41	-60.92 – 0.09	0.052
Flat	-14.96	-29.44 – -0.48	0.044
Undeveloped	27.03	11.08 – 42.98	0.001
Riskofcontamination 10	-0.80	-3.15 – 1.55	0.505
Observations	288
R² / adjusted R²	0.246 / 0.219

## [1] "########## LOGISTIC ##########"

	TTC 0
Predictors	Odds Ratios	CI	p
(Intercept)	1.27	0.52 – 3.09	0.606
`_4_18_enoughfundsrepair`	1.99	1.08 – 3.68	0.027
Borehole mech pump	1.04	0.37 – 2.91	0.945
Dug well open	4.89	1.58 – 15.10	0.006
Dug well handpump	2.13	1.01 – 4.50	0.047
Spring protected	3.90	1.38 – 11.06	0.010
Spring unprotected	5261192.76	0.00 – Inf	0.984
Elevated	0.48	0.12 – 1.93	0.303
Flat	0.68	0.34 – 1.36	0.275
Undeveloped	3.58	1.44 – 8.90	0.006
Riskofcontamination 10	0.94	0.84 – 1.06	0.325
Observations	288
Cox & Snell’s R² / Nagelkerke’s R²	0.157 / 0.221

## [1] "########## ZINBR ##########"

	TTC Round
Predictors	Incidence Rate Ratios	CI	p
(Intercept)	70.34	66.13 – 74.83	<0.001
`_4_18_enoughfundsrepair`	0.73	0.69 – 0.76	<0.001
Borehole mech pump	1.34	1.23 – 1.47	<0.001
Dug well open	1.14	1.07 – 1.22	<0.001
Dug well handpump	0.86	0.82 – 0.91	<0.001
Spring protected	0.73	0.68 – 0.78	<0.001
Spring unprotected	1.16	1.07 – 1.26	<0.001
Elevated	0.35	0.29 – 0.43	<0.001
Flat	0.79	0.75 – 0.83	<0.001
Undeveloped	1.24	1.18 – 1.30	<0.001
Riskofcontamination 10	1.00	1.00 – 1.01	0.309
Zero-Inflated Model
(Intercept)	0.84	0.35 – 2.04	0.708
`_4_18_enoughfundsrepair`	0.54	0.29 – 0.98	0.044
Borehole mech pump	0.81	0.29 – 2.28	0.695
Dug well open	0.23	0.08 – 0.68	0.008
Dug well handpump	0.42	0.20 – 0.89	0.023
Spring protected	0.28	0.10 – 0.77	0.014
Spring unprotected	0.00	0.00 – Inf	0.990
Elevated	1.86	0.47 – 7.40	0.380
Flat	1.58	0.80 – 3.14	0.192
Undeveloped	0.25	0.10 – 0.63	0.003
Riskofcontamination 10	1.06	0.95 – 1.19	0.302

## Number of observations in ZINBR regression above:  288

5.7 _4_19_personrepair as IV

## [1] "########## OLS ##########"

	TTC
Predictors	Estimates	CI	p
(Intercept)	42.00	23.15 – 60.86	<0.001
`_4_19_personrepair`	-1.48	-14.07 – 11.11	0.818
Borehole mech pump	10.37	-12.65 – 33.39	0.378
Dug well open	32.66	12.95 – 52.38	0.001
Dug well handpump	5.83	-8.96 – 20.62	0.440
Spring protected	-1.10	-21.13 – 18.93	0.914
Spring unprotected	42.08	9.10 – 75.06	0.013
Elevated	-33.62	-62.97 – -4.26	0.026
Flat	-17.09	-31.64 – -2.54	0.022
Undeveloped	27.02	11.05 – 42.99	0.001
Riskofcontamination 10	-0.85	-3.24 – 1.54	0.485
Observations	292
R² / adjusted R²	0.250 / 0.223

## [1] "########## LOGISTIC ##########"

	TTC 0
Predictors	Odds Ratios	CI	p
(Intercept)	1.54	0.63 – 3.76	0.345
`_4_19_personrepair`	1.19	0.65 – 2.16	0.577
Borehole mech pump	1.30	0.47 – 3.59	0.614
Dug well open	3.87	1.34 – 11.24	0.013
Dug well handpump	2.37	1.14 – 4.92	0.021
Spring protected	3.20	1.14 – 9.01	0.027
Spring unprotected	4663800.64	0.00 – Inf	0.984
Elevated	0.57	0.15 – 2.21	0.415
Flat	0.63	0.31 – 1.25	0.185
Undeveloped	2.99	1.22 – 7.35	0.017
Riskofcontamination 10	0.95	0.85 – 1.07	0.433
Observations	292
Cox & Snell’s R² / Nagelkerke’s R²	0.151 / 0.211

## [1] "########## ZINBR ##########"

	TTC Round
Predictors	Incidence Rate Ratios	CI	p
(Intercept)	67.37	63.20 – 71.80	<0.001
`_4_19_personrepair`	0.92	0.88 – 0.96	<0.001
Borehole mech pump	1.18	1.09 – 1.29	<0.001
Dug well open	1.23	1.16 – 1.31	<0.001
Dug well handpump	0.90	0.85 – 0.95	<0.001
Spring protected	0.72	0.67 – 0.78	<0.001
Spring unprotected	1.23	1.13 – 1.33	<0.001
Elevated	0.30	0.25 – 0.36	<0.001
Flat	0.78	0.74 – 0.82	<0.001
Undeveloped	1.31	1.25 – 1.38	<0.001
Riskofcontamination 10	1.00	0.99 – 1.01	0.423
Zero-Inflated Model
(Intercept)	0.69	0.29 – 1.68	0.418
`_4_19_personrepair`	0.91	0.50 – 1.65	0.755
Borehole mech pump	0.65	0.24 – 1.81	0.415
Dug well open	0.28	0.10 – 0.79	0.017
Dug well handpump	0.38	0.18 – 0.79	0.009
Spring protected	0.34	0.13 – 0.94	0.038
Spring unprotected	0.00	0.00 – Inf	0.990
Elevated	1.56	0.40 – 6.04	0.522
Flat	1.69	0.85 – 3.35	0.132
Undeveloped	0.30	0.12 – 0.74	0.009
Riskofcontamination 10	1.05	0.94 – 1.18	0.402

## Number of observations in ZINBR regression above:  292

5.8 _4_20_persontrained as IV

## [1] "########## OLS ##########"

	TTC
Predictors	Estimates	CI	p
(Intercept)	42.78	25.55 – 60.02	<0.001
`_4_20_persontrained`	1.30	-15.91 – 18.52	0.882
Borehole mech pump	9.85	-13.20 – 32.91	0.403
Dug well open	31.55	12.35 – 50.76	0.001
Dug well handpump	6.29	-8.44 – 21.02	0.403
Spring protected	-0.02	-19.01 – 18.96	0.998
Spring unprotected	40.70	7.90 – 73.50	0.016
Elevated	-36.03	-65.21 – -6.85	0.016
Flat	-19.36	-33.71 – -5.01	0.009
Undeveloped	26.06	10.42 – 41.69	0.001
Riskofcontamination 10	-0.67	-3.00 – 1.66	0.572
Observations	296
R² / adjusted R²	0.248 / 0.222

## [1] "########## LOGISTIC ##########"

	TTC 0
Predictors	Odds Ratios	CI	p
(Intercept)	2.01	0.87 – 4.63	0.102
`_4_20_persontrained`	0.88	0.40 – 1.95	0.759
Borehole mech pump	1.25	0.45 – 3.46	0.671
Dug well open	3.41	1.21 – 9.62	0.021
Dug well handpump	2.06	1.00 – 4.25	0.049
Spring protected	2.94	1.08 – 7.95	0.034
Spring unprotected	4136364.24	0.00 – Inf	0.984
Elevated	0.53	0.14 – 2.03	0.352
Flat	0.60	0.30 – 1.18	0.139
Undeveloped	2.85	1.18 – 6.91	0.020
Riskofcontamination 10	0.95	0.84 – 1.06	0.334
Observations	296
Cox & Snell’s R² / Nagelkerke’s R²	0.143 / 0.201

## [1] "########## ZINBR ##########"

	TTC Round
Predictors	Incidence Rate Ratios	CI	p
(Intercept)	63.04	59.56 – 66.73	<0.001
`_4_20_persontrained`	1.08	1.02 – 1.15	0.011
Borehole mech pump	1.17	1.07 – 1.27	<0.001
Dug well open	1.25	1.18 – 1.33	<0.001
Dug well handpump	0.94	0.89 – 0.99	0.012
Spring protected	0.77	0.72 – 0.83	<0.001
Spring unprotected	1.24	1.14 – 1.34	<0.001
Elevated	0.29	0.25 – 0.35	<0.001
Flat	0.76	0.72 – 0.79	<0.001
Undeveloped	1.31	1.25 – 1.37	<0.001
Riskofcontamination 10	1.00	0.99 – 1.01	0.479
Zero-Inflated Model
(Intercept)	0.55	0.24 – 1.25	0.153
`_4_20_persontrained`	1.17	0.53 – 2.57	0.705
Borehole mech pump	0.69	0.25 – 1.91	0.475
Dug well open	0.31	0.11 – 0.86	0.024
Dug well handpump	0.43	0.21 – 0.89	0.022
Spring protected	0.36	0.14 – 0.96	0.040
Spring unprotected	0.00	0.00 – Inf	0.990
Elevated	1.72	0.45 – 6.60	0.429
Flat	1.77	0.90 – 3.48	0.098
Undeveloped	0.31	0.13 – 0.75	0.009
Riskofcontamination 10	1.06	0.95 – 1.19	0.301

## Number of observations in ZINBR regression above:  296

5.9 _5_0_latrineaccess as IV

## [1] "########## OLS ##########"

	TTC
Predictors	Estimates	CI	p
(Intercept)	60.51	15.46 – 105.57	0.009
`_5_0_latrineaccess`	-0.17	-0.61 – 0.26	0.430
Borehole mech pump	11.14	-12.42 – 34.70	0.355
Dug well open	32.37	13.07 – 51.67	0.001
Dug well handpump	5.30	-9.45 – 20.05	0.482
Spring protected	0.04	-19.56 – 19.64	0.997
Spring unprotected	41.46	8.44 – 74.48	0.014
Elevated	-36.14	-65.47 – -6.80	0.016
Flat	-19.40	-33.87 – -4.92	0.009
Undeveloped	24.44	8.78 – 40.09	0.002
Riskofcontamination 10	-0.86	-3.21 – 1.49	0.474
Observations	294
R² / adjusted R²	0.241 / 0.214

## [1] "########## LOGISTIC ##########"

	TTC 0
Predictors	Odds Ratios	CI	p
(Intercept)	1.96	0.24 – 16.18	0.532
`_5_0_latrineaccess`	1.00	0.98 – 1.02	0.930
Borehole mech pump	1.32	0.47 – 3.76	0.599
Dug well open	3.61	1.28 – 10.21	0.015
Dug well handpump	2.10	1.02 – 4.33	0.044
Spring protected	2.85	1.04 – 7.84	0.042
Spring unprotected	4594445.18	0.00 – Inf	0.984
Elevated	0.52	0.13 – 1.98	0.336
Flat	0.60	0.30 – 1.20	0.148
Undeveloped	2.62	1.11 – 6.20	0.028
Riskofcontamination 10	0.93	0.83 – 1.05	0.242
Observations	294
Cox & Snell’s R² / Nagelkerke’s R²	0.134 / 0.189

## [1] "########## ZINBR ##########"

	TTC Round
Predictors	Incidence Rate Ratios	CI	p
(Intercept)	100.37	87.64 – 114.94	<0.001
`_5_0_latrineaccess`	1.00	0.99 – 1.00	<0.001
Borehole mech pump	1.16	1.07 – 1.27	<0.001
Dug well open	1.22	1.15 – 1.30	<0.001
Dug well handpump	0.92	0.87 – 0.96	0.001
Spring protected	0.76	0.71 – 0.82	<0.001
Spring unprotected	1.20	1.11 – 1.30	<0.001
Elevated	0.29	0.24 – 0.35	<0.001
Flat	0.74	0.71 – 0.78	<0.001
Undeveloped	1.28	1.22 – 1.33	<0.001
Riskofcontamination 10	1.01	1.00 – 1.01	0.130
Zero-Inflated Model
(Intercept)	0.62	0.08 – 5.03	0.651
`_5_0_latrineaccess`	1.00	0.98 – 1.02	0.858
Borehole mech pump	0.65	0.23 – 1.85	0.419
Dug well open	0.29	0.11 – 0.80	0.017
Dug well handpump	0.42	0.20 – 0.87	0.020
Spring protected	0.38	0.14 – 1.00	0.051
Spring unprotected	0.00	0.00 – Inf	0.990
Elevated	1.75	0.46 – 6.70	0.411
Flat	1.76	0.90 – 3.47	0.100
Undeveloped	0.34	0.15 – 0.80	0.014
Riskofcontamination 10	1.07	0.96 – 1.20	0.219

## Number of observations in ZINBR regression above:  294

5.10 _5_1_a_facilitator as IV

## [1] "########## OLS ##########"

	TTC
Predictors	Estimates	CI	p
(Intercept)	42.69	23.28 – 62.10	<0.001
`_5_1_a_facilitator`	5.95	-6.09 – 17.99	0.334
Borehole mech pump	1.16	-22.64 – 24.97	0.924
Dug well open	36.18	15.73 – 56.62	0.001
Dug well handpump	0.45	-15.21 – 16.10	0.956
Spring protected	0.11	-20.50 – 20.72	0.992
Spring unprotected	41.27	7.78 – 74.76	0.016
Elevated	-33.91	-63.51 – -4.31	0.026
Flat	-15.49	-30.70 – -0.28	0.047
Undeveloped	27.53	11.02 – 44.04	0.001
Riskofcontamination 10	-1.43	-3.98 – 1.11	0.272
Observations	264
R² / adjusted R²	0.243 / 0.213

## [1] "########## LOGISTIC ##########"

	TTC 0
Predictors	Odds Ratios	CI	p
(Intercept)	1.98	0.78 – 4.98	0.148
`_5_1_a_facilitator`	0.92	0.51 – 1.65	0.781
Borehole mech pump	0.95	0.33 – 2.69	0.922
Dug well open	3.94	1.29 – 12.02	0.016
Dug well handpump	1.90	0.89 – 4.08	0.098
Spring protected	3.13	1.07 – 9.20	0.038
Spring unprotected	4909205.48	0.00 – Inf	0.984
Elevated	0.60	0.15 – 2.35	0.463
Flat	0.77	0.38 – 1.58	0.479
Undeveloped	3.01	1.22 – 7.43	0.017
Riskofcontamination 10	0.92	0.81 – 1.05	0.209
Observations	264
Cox & Snell’s R² / Nagelkerke’s R²	0.133 / 0.187

## [1] "########## ZINBR ##########"

	TTC Round
Predictors	Incidence Rate Ratios	CI	p
(Intercept)	61.38	57.48 – 65.54	<0.001
`_5_1_a_facilitator`	1.21	1.16 – 1.26	<0.001
Borehole mech pump	1.03	0.94 – 1.14	0.471
Dug well open	1.31	1.23 – 1.40	<0.001
Dug well handpump	0.86	0.82 – 0.91	<0.001
Spring protected	0.76	0.71 – 0.82	<0.001
Spring unprotected	1.19	1.09 – 1.29	<0.001
Elevated	0.30	0.25 – 0.36	<0.001
Flat	0.79	0.75 – 0.83	<0.001
Undeveloped	1.33	1.26 – 1.39	<0.001
Riskofcontamination 10	0.99	0.99 – 1.00	0.261
Zero-Inflated Model
(Intercept)	0.51	0.20 – 1.28	0.151
`_5_1_a_facilitator`	1.35	0.76 – 2.40	0.311
Borehole mech pump	0.90	0.32 – 2.56	0.842
Dug well open	0.28	0.09 – 0.82	0.020
Dug well handpump	0.46	0.22 – 0.99	0.048
Spring protected	0.34	0.12 – 0.97	0.044
Spring unprotected	0.00	0.00 – Inf	0.990
Elevated	1.55	0.40 – 6.02	0.529
Flat	1.38	0.68 – 2.80	0.368
Undeveloped	0.30	0.12 – 0.74	0.009
Riskofcontamination 10	1.09	0.96 – 1.23	0.176

## Number of observations in ZINBR regression above:  264

## [1] "########## ORDINAL LOGISTIC ##########"

## Warning in polr(TTCgrp50 ~ `_5_1_a_facilitator` + Type + Topo + Land +
## Riskofcontamination10, : design appears to be rank-deficient, so dropping
## some coefs

	TT Cgrp 50
Predictors	Odds Ratios	CI	p
`_5_1_a_facilitator`	1.38	0.76 – 2.53	0.294
Borehole mech pump	0.82	0.21 – 3.23	0.777
Dug well open	4.11	1.54 – 10.94	0.005
Dug well handpump	0.89	0.41 – 1.92	0.760
Spring protected	0.85	0.30 – 2.46	0.767
Spring unprotected	5.13	0.89 – 29.42	0.067
Elevated	0.00	0.00 – 0.00	<0.001
Flat	0.48	0.23 – 1.00	0.050
Undeveloped	2.53	1.20 – 5.33	0.014
Riskofcontamination 10	0.96	0.85 – 1.09	0.555
(Intercept: (-Inf,50]\|(50,100])	2.18	0.86 – 5.55	0.102
(Intercept: (50,100]\|(100,150])	3.15	1.23 – 8.09	0.017
Observations	264
Cox & Snell’s R² / Nagelkerke’s R²	0.458 / 0.533

6 Building a regression model – All variables together

6.1 Zero-inflated negative-binomial regression

Without specifying which variable(s) to inflate:

	TTC Round
Predictors	Incidence Rate Ratios	CI	p
(Intercept)	115.70	97.12 – 137.83	<0.001
`_4_3_a_authorizedlist`	0.76	0.71 – 0.80	<0.001
`_4_4_feecollectionsystem`	0.90	0.86 – 0.95	<0.001
`_4_17_planOandM`	1.02	0.98 – 1.07	0.387
`_4_18_enoughfundsrepair`	0.79	0.75 – 0.84	<0.001
`_4_19_personrepair`	0.89	0.83 – 0.94	<0.001
`_4_20_persontrained`	1.04	0.96 – 1.12	0.365
`_5_0_latrineaccess`	1.00	0.99 – 1.00	<0.001
`_5_1_a_facilitator`	1.25	1.19 – 1.30	<0.001
Borehole mech pump	1.07	0.96 – 1.18	0.215
Dug well open	0.98	0.90 – 1.06	0.587
Dug well handpump	0.79	0.74 – 0.84	<0.001
Spring protected	0.57	0.52 – 0.63	<0.001
Spring unprotected	0.96	0.87 – 1.05	0.375
Elevated	0.35	0.29 – 0.44	<0.001
Flat	0.79	0.75 – 0.84	<0.001
Undeveloped	1.28	1.21 – 1.36	<0.001
Riskofcontamination 10	1.00	0.99 – 1.01	0.860
Zero-Inflated Model
(Intercept)	0.17	0.01 – 2.37	0.189
`_4_3_a_authorizedlist`	1.35	0.64 – 2.85	0.438
`_4_4_feecollectionsystem`	0.96	0.46 – 2.01	0.910
`_4_17_planOandM`	1.28	0.60 – 2.73	0.524
`_4_18_enoughfundsrepair`	0.43	0.20 – 0.92	0.030
`_4_19_personrepair`	0.97	0.44 – 2.15	0.947
`_4_20_persontrained`	0.71	0.25 – 2.00	0.519
`_5_0_latrineaccess`	1.01	0.99 – 1.04	0.306
`_5_1_a_facilitator`	1.49	0.78 – 2.85	0.226
Borehole mech pump	1.02	0.31 – 3.33	0.973
Dug well open	0.25	0.06 – 1.00	0.050
Dug well handpump	0.42	0.18 – 1.01	0.052
Spring protected	0.32	0.09 – 1.13	0.077
Spring unprotected	0.00	0.00 – Inf	0.990
Elevated	1.66	0.35 – 7.87	0.523
Flat	1.67	0.76 – 3.67	0.201
Undeveloped	0.26	0.09 – 0.72	0.010
Riskofcontamination 10	1.09	0.95 – 1.25	0.221

## Number of observations in ZINBR regression above:  228

Specifying inflation of water source type:

	TTC Round
Predictors	Incidence Rate Ratios	CI	p
(Intercept)	127.38	107.39 – 151.10	<0.001
`_4_3_a_authorizedlist`	0.74	0.70 – 0.79	<0.001
`_4_4_feecollectionsystem`	0.90	0.85 – 0.94	<0.001
`_4_17_planOandM`	1.00	0.95 – 1.04	0.883
`_4_18_enoughfundsrepair`	0.80	0.76 – 0.84	<0.001
`_4_19_personrepair`	0.91	0.85 – 0.96	0.001
`_4_20_persontrained`	1.01	0.93 – 1.08	0.875
`_5_0_latrineaccess`	0.99	0.99 – 1.00	<0.001
`_5_1_a_facilitator`	1.25	1.20 – 1.30	<0.001
Elevated	0.30	0.25 – 0.37	<0.001
Flat	0.79	0.74 – 0.83	<0.001
Undeveloped	1.27	1.21 – 1.35	<0.001
Riskofcontamination 10	0.99	0.98 – 1.00	0.021
Zero-Inflated Model
(Intercept)	0.90	0.61 – 1.34	0.612
Borehole mech pump	0.78	0.27 – 2.21	0.634
Dug well open	0.24	0.09 – 0.63	0.004
Dug well handpump	0.35	0.16 – 0.76	0.009
Spring protected	0.43	0.17 – 1.13	0.086
Spring unprotected	0.00	0.00 – Inf	0.990

## Number of observations in ZINBR regression above:  228

6.2 Ordinal Logistic Regression and Negative Binomial Regression

This is just an experiment for now and needs more work. For this regression, TTC has been divided into three categories, as the table below shows:

##            
##             Count <NA>
##   (-Inf,50]   230    0
##   (50,100]     20    0
##   (100,150]    84    0
##   <NA>         20    0

And here is an ordinal logistic regression with this grouped TTC variable as the dependent variable:

olog1 <- polr(TTCgrp50 ~ `_4_3_a_authorizedlist` + `_4_4_feecollectionsystem` + `_4_17_planOandM` + `_4_18_enoughfundsrepair` + `_4_19_personrepair` + `_4_20_persontrained` + `_5_0_latrineaccess` + `_5_1_a_facilitator` + Type + Topo + Land + Riskofcontamination10, data = db, Hess=TRUE)

## Warning in polr(TTCgrp50 ~ `_4_3_a_authorizedlist` +
## `_4_4_feecollectionsystem` + : design appears to be rank-deficient, so
## dropping some coefs

tab_model(olog1)

	TT Cgrp 50
Predictors	Odds Ratios	CI	p
`_4_3_a_authorizedlist`	0.47	0.19 – 1.13	0.092
`_4_4_feecollectionsystem`	0.81	0.37 – 1.77	0.590
`_4_17_planOandM`	0.92	0.45 – 1.90	0.827
`_4_18_enoughfundsrepair`	0.78	0.36 – 1.73	0.546
`_4_19_personrepair`	0.65	0.27 – 1.55	0.327
`_4_20_persontrained`	1.43	0.47 – 4.35	0.523
`_5_0_latrineaccess`	0.99	0.97 – 1.01	0.404
`_5_1_a_facilitator`	1.50	0.76 – 2.99	0.245
Borehole mech pump	0.76	0.18 – 3.24	0.712
Dug well open	2.11	0.60 – 7.45	0.246
Dug well handpump	0.73	0.31 – 1.75	0.483
Spring protected	0.47	0.13 – 1.71	0.255
Spring unprotected	5.06	0.47 – 53.93	0.179
Elevated	0.00	0.00 – 0.00	<0.001
Flat	0.47	0.21 – 1.06	0.069
Undeveloped	2.61	1.11 – 6.11	0.027
Riskofcontamination 10	0.97	0.84 – 1.13	0.711
(Intercept: (-Inf,50]\|(50,100])	0.44	0.03 – 5.99	0.537
(Intercept: (50,100]\|(100,150])	0.61	0.05 – 8.38	0.715
Observations	228
Cox & Snell’s R² / Nagelkerke’s R²	0.627 / 0.700

Negative Binomial Regression for the same DV:

nbTTC <- glm.nb(as.numeric(TTCgrp50) ~ `_4_3_a_authorizedlist` + `_4_4_feecollectionsystem` + `_4_17_planOandM` + `_4_18_enoughfundsrepair` + `_4_19_personrepair` + `_4_20_persontrained` + `_5_0_latrineaccess` + `_5_1_a_facilitator` + Type + Topo + Land + Riskofcontamination10, data = db)

## Warning in theta.ml(Y, mu, sum(w), w, limit = control$maxit, trace =
## control$trace > : iteration limit reached

## Warning in theta.ml(Y, mu, sum(w), w, limit = control$maxit, trace =
## control$trace > : iteration limit reached

tab_model(nbTTC)

	as numeric(TT Cgrp 50)
Predictors	Incidence Rate Ratios	CI	p
(Intercept)	2.14	0.87 – 5.26	0.097
`_4_3_a_authorizedlist`	0.87	0.66 – 1.15	0.337
`_4_4_feecollectionsystem`	0.94	0.72 – 1.21	0.617
`_4_17_planOandM`	0.99	0.78 – 1.26	0.955
`_4_18_enoughfundsrepair`	0.95	0.73 – 1.24	0.704
`_4_19_personrepair`	0.92	0.69 – 1.23	0.579
`_4_20_persontrained`	1.08	0.75 – 1.56	0.689
`_5_0_latrineaccess`	1.00	0.99 – 1.01	0.659
`_5_1_a_facilitator`	1.08	0.87 – 1.36	0.484
Borehole mech pump	0.94	0.59 – 1.51	0.807
Dug well open	1.19	0.78 – 1.80	0.421
Dug well handpump	0.92	0.69 – 1.25	0.609
Spring protected	0.83	0.54 – 1.28	0.394
Spring unprotected	1.24	0.72 – 2.13	0.448
Elevated	0.67	0.33 – 1.35	0.262
Flat	0.86	0.64 – 1.14	0.295
Undeveloped	1.28	0.94 – 1.73	0.112
Riskofcontamination 10	0.99	0.95 – 1.04	0.753
Observations	228
Cox & Snell’s R² / Nagelkerke’s R²	0.119 / 0.327

6.3 Standard Logistic Regression

This section contains logistic regressions for a number of dependent variables.

DV = TTC0:

	TTC 0
Predictors	Odds Ratios	CI	p
(Intercept)	6.30	0.44 – 89.93	0.175
`_4_3_a_authorizedlist`	0.62	0.29 – 1.31	0.213
`_4_4_feecollectionsystem`	0.96	0.45 – 2.02	0.905
`_4_17_planOandM`	0.82	0.38 – 1.76	0.612
`_4_18_enoughfundsrepair`	2.45	1.13 – 5.31	0.023
`_4_19_personrepair`	1.07	0.48 – 2.38	0.864
`_4_20_persontrained`	1.28	0.45 – 3.67	0.644
`_5_0_latrineaccess`	0.99	0.96 – 1.01	0.300
`_5_1_a_facilitator`	0.87	0.45 – 1.68	0.679
Borehole mech pump	0.75	0.23 – 2.45	0.634
Dug well open	4.14	0.97 – 17.71	0.055
Dug well handpump	2.04	0.85 – 4.86	0.110
Spring protected	3.19	0.86 – 11.81	0.082
Spring unprotected	4469089.65	0.00 – Inf	0.985
Elevated	0.52	0.11 – 2.52	0.415
Flat	0.64	0.29 – 1.44	0.280
Undeveloped	3.57	1.25 – 10.19	0.018
Riskofcontamination 10	0.93	0.81 – 1.07	0.291
Observations	228
Cox & Snell’s R² / Nagelkerke’s R²	0.165 / 0.232

DV = TTCGE1:

	db$TTCGE 1
Predictors	Odds Ratios	CI	p
(Intercept)	5.77	0.42 – 79.04	0.189
`_4_3_a_authorizedlist`	0.74	0.35 – 1.57	0.438
`_4_4_feecollectionsystem`	1.04	0.50 – 2.19	0.910
`_4_17_planOandM`	0.78	0.37 – 1.67	0.524
`_4_18_enoughfundsrepair`	2.32	1.08 – 4.96	0.030
`_4_19_personrepair`	1.03	0.46 – 2.27	0.948
`_4_20_persontrained`	1.40	0.50 – 3.94	0.518
`_5_0_latrineaccess`	0.99	0.96 – 1.01	0.305
`_5_1_a_facilitator`	0.67	0.35 – 1.28	0.226
Borehole mech pump	0.98	0.30 – 3.20	0.973
Dug well open	4.07	1.00 – 16.55	0.050
Dug well handpump	2.37	0.99 – 5.66	0.052
Spring protected	3.17	0.88 – 11.41	0.077
Spring unprotected	5381982.69	0.00 – Inf	0.984
Elevated	0.60	0.13 – 2.85	0.523
Flat	0.60	0.27 – 1.31	0.201
Undeveloped	3.91	1.39 – 10.99	0.010
Riskofcontamination 10	0.92	0.80 – 1.05	0.221
Observations	228
Cox & Snell’s R² / Nagelkerke’s R²	0.182 / 0.252

DV = TTCGE10:

	db$TTCGE 10
Predictors	Odds Ratios	CI	p
(Intercept)	16.14	1.26 – 205.90	0.032
`_4_3_a_authorizedlist`	0.57	0.27 – 1.20	0.139
`_4_4_feecollectionsystem`	0.92	0.45 – 1.88	0.820
`_4_17_planOandM`	0.79	0.39 – 1.57	0.498
`_4_18_enoughfundsrepair`	1.87	0.90 – 3.86	0.093
`_4_19_personrepair`	0.79	0.37 – 1.71	0.553
`_4_20_persontrained`	1.48	0.54 – 4.03	0.447
`_5_0_latrineaccess`	0.98	0.95 – 1.00	0.050
`_5_1_a_facilitator`	0.96	0.52 – 1.80	0.907
Borehole mech pump	0.89	0.27 – 2.93	0.847
Dug well open	2.06	0.60 – 7.06	0.253
Dug well handpump	1.52	0.69 – 3.37	0.302
Spring protected	1.68	0.53 – 5.30	0.376
Spring unprotected	3.79	0.36 – 40.18	0.268
Elevated	0.34	0.07 – 1.69	0.187
Flat	0.35	0.16 – 0.74	0.006
Undeveloped	1.83	0.77 – 4.36	0.173
Riskofcontamination 10	0.94	0.83 – 1.07	0.380
Observations	228
Cox & Snell’s R² / Nagelkerke’s R²	0.172 / 0.229

DV = TTCGE50:

	TTCGE 50
Predictors	Odds Ratios	CI	p
(Intercept)	4.26	0.30 – 60.49	0.284
`_4_3_a_authorizedlist`	0.44	0.18 – 1.10	0.079
`_4_4_feecollectionsystem`	0.78	0.34 – 1.76	0.547
`_4_17_planOandM`	0.82	0.39 – 1.72	0.592
`_4_18_enoughfundsrepair`	0.81	0.37 – 1.79	0.601
`_4_19_personrepair`	0.65	0.27 – 1.58	0.342
`_4_20_persontrained`	1.29	0.42 – 3.93	0.656
`_5_0_latrineaccess`	0.99	0.96 – 1.01	0.293
`_5_1_a_facilitator`	1.59	0.78 – 3.23	0.200
Borehole mech pump	0.97	0.25 – 3.75	0.961
Dug well open	2.14	0.58 – 7.84	0.253
Dug well handpump	0.76	0.31 – 1.84	0.538
Spring protected	0.50	0.13 – 1.85	0.299
Spring unprotected	3.63	0.33 – 39.75	0.290
Elevated	0.00	0.00 – Inf	0.983
Flat	0.41	0.18 – 0.95	0.037
Undeveloped	2.61	1.07 – 6.35	0.034
Riskofcontamination 10	0.93	0.80 – 1.08	0.335
Observations	228
Cox & Snell’s R² / Nagelkerke’s R²	0.243 / 0.336

DV = TTCGE100:

	TTCGE 100
Predictors	Odds Ratios	CI	p
(Intercept)	0.69	0.04 – 12.29	0.804
`_4_3_a_authorizedlist`	0.63	0.24 – 1.65	0.347
`_4_4_feecollectionsystem`	0.67	0.28 – 1.58	0.356
`_4_17_planOandM`	1.14	0.52 – 2.51	0.749
`_4_18_enoughfundsrepair`	0.81	0.35 – 1.92	0.640
`_4_19_personrepair`	0.62	0.24 – 1.63	0.331
`_4_20_persontrained`	1.66	0.50 – 5.52	0.411
`_5_0_latrineaccess`	1.00	0.97 – 1.02	0.725
`_5_1_a_facilitator`	1.37	0.65 – 2.89	0.407
Borehole mech pump	0.87	0.20 – 3.81	0.857
Dug well open	2.38	0.63 – 8.93	0.199
Dug well handpump	0.65	0.25 – 1.70	0.384
Spring protected	0.37	0.09 – 1.54	0.169
Spring unprotected	5.58	0.51 – 61.25	0.159
Elevated	0.00	0.00 – Inf	0.990
Flat	0.65	0.26 – 1.61	0.348
Undeveloped	3.31	1.28 – 8.57	0.013
Riskofcontamination 10	0.98	0.83 – 1.15	0.791
Observations	228
Cox & Snell’s R² / Nagelkerke’s R²	0.235 / 0.337

DV = TTCGE125:

	TTCGE 125
Predictors	Odds Ratios	CI	p
(Intercept)	0.78	0.04 – 14.17	0.865
`_4_3_a_authorizedlist`	0.66	0.25 – 1.75	0.402
`_4_4_feecollectionsystem`	0.59	0.24 – 1.44	0.245
`_4_17_planOandM`	1.01	0.45 – 2.26	0.985
`_4_18_enoughfundsrepair`	0.63	0.26 – 1.54	0.312
`_4_19_personrepair`	0.91	0.34 – 2.43	0.854
`_4_20_persontrained`	1.14	0.33 – 3.96	0.841
`_5_0_latrineaccess`	0.99	0.97 – 1.02	0.665
`_5_1_a_facilitator`	1.44	0.67 – 3.10	0.350
Borehole mech pump	1.16	0.26 – 5.18	0.845
Dug well open	2.13	0.55 – 8.17	0.272
Dug well handpump	0.57	0.21 – 1.53	0.265
Spring protected	0.35	0.08 – 1.50	0.158
Spring unprotected	4.46	0.40 – 49.64	0.224
Elevated	0.00	0.00 – Inf	0.990
Flat	0.56	0.22 – 1.41	0.218
Undeveloped	3.41	1.31 – 8.85	0.012
Riskofcontamination 10	1.00	0.85 – 1.19	0.965
Observations	228
Cox & Snell’s R² / Nagelkerke’s R²	0.246 / 0.355

DV=0 if TTC=0; DV=1 if TTC=125; DV = missing otherwise

	TTC 0 or 125
Predictors	Odds Ratios	CI	p
(Intercept)	1.88	0.03 – 129.60	0.769
`_4_3_a_authorizedlist`	0.58	0.17 – 1.94	0.376
`_4_4_feecollectionsystem`	0.51	0.16 – 1.63	0.254
`_4_17_planOandM`	0.95	0.31 – 2.90	0.928
`_4_18_enoughfundsrepair`	0.96	0.28 – 3.29	0.952
`_4_19_personrepair`	1.23	0.34 – 4.50	0.755
`_4_20_persontrained`	1.13	0.22 – 5.95	0.883
`_5_0_latrineaccess`	1.00	0.96 – 1.04	0.809
`_5_1_a_facilitator`	1.22	0.44 – 3.35	0.702
Borehole mech pump	0.92	0.17 – 4.88	0.925
Dug well open	6.56	1.01 – 42.64	0.049
Dug well handpump	1.13	0.30 – 4.19	0.859
Spring protected	1.16	0.19 – 7.04	0.868
Spring unprotected	11795902.99	0.00 – Inf	0.991
Elevated	0.00	0.00 – Inf	0.992
Flat	0.44	0.14 – 1.43	0.173
Undeveloped	6.48	1.81 – 23.27	0.004
Riskofcontamination 10	0.90	0.74 – 1.10	0.298
Observations	134
Cox & Snell’s R² / Nagelkerke’s R²	0.370 / 0.493

TTC0 is a binary variable in which 0 means that TTC=0 and 1 means that TTC>0↩
For ZINBR regressions, TTC has been rounded to the nearest integer value↩

WHA Gulu Survey Analysis for H1