Test Yourself

Question #1:

None of the interactions are significant, therefore they shouldn't be included in the equation.

 Regression Table: Question: Does Drinking water source have an impact on weight/age z score controlling for SES (roof) and education? Dependent Variable: WAZ Independent Variable Model Number: coefficient (t,p) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Dummy Good Drinking water 0.314 (2.252, 0.025) - 0.258 (1.794, 0.073) -0.310 (-0.682, 0.495) 0.244 (1.687, 0.092) -0.279 (-0.611, 0.541) - 0.178 (1.219, 0.223) - 0.163 (1.115, 0.265) 0.194 (1.346, 0.179) 0.233 (0.855, 0.393) 0.273 (1.936, 0.053) 0.03848 (0.200, 0.842) Dummy Med. Drinking water 0.005133 (0.049, 0.961) - -0.04294 (-0.395, 0.693) -0.03389 (-0.312, 0.755) 0.162 (0.615, 0.539) 0.128 (0.486, 0.627) - -0.0852 (-0.777, 0.438) - -0.0864 (-0.788, 0.431) -0.0716 (-0.661, 0.509) 0.112 (0.689, 0.491) -0.00994 (-0.094, 0.925) -0.0414 (-0.317, 0.751) Toilet Access - 0.242 (2.117, 0.035) 0.224 (1.876, 0.061) 0.181 (1.472, 0.141) 0.278 (2.058, 0.040) 0.228 (1.613, 0.107) - 0.08540 (0.663, 0.507) - 0.06814 (0.526, 0.599) - - - - Interaction (dgooddrwat *toilet) - - - 0.627 (1.318, 0.188) - 0.581 (1.207, 0.228) - - - - - - - - Interaction (dmeddrwat *toilet) - - - - -0.245 (-0.855, 0.393) -0.195 (-0.674, 0.501) - - - - - - - - Roof - - - - - - 0.359 (3.786, 0.000) 0.304 (2.801, 0.005) - 0.283 (2.572, 0.010) 0.323 (3.231, 0.001) 0.448 (3.312, 0.001) - - Interaction (dgooddrwat * roof) - - - - - - - - - - - -0.111 (-0.344, 0.731) - - Interaction (dmeddrwat * roof) - - - - - - - - - - - -0.331 (-1.516, 0.130) - - Educational Attainment - - - - - - - - 0.267 (2.733, 0.006) 0.122 (1.180, 0.238) - - 0.195 (1.940, 0.053) 0.07095 (0.489, 0.625) Interaction (dgooddrwat * edu) - - - - - - - - - - - - - 0.509 (1.787, 0.074) Interaction (dmeddrwat * edu) - - - - - - - - - - - - - 0.109 (0.488, 0.626) N 775 803 773 773 773 773 802 770 806 770 772 772 775 775 Adjusted R2 0.005 0.004 0.008 0.009 0.007 0.008 0.016 0.016 0.008 0.017 0.017 0.017 0.008 0.01

Question #2:

 Regression Table Question: Does delivery location have an affect on weight/age z score controlling for SES and education? Dependent Variable: WAZ Independent Variable Model Number: coefficient (t,p) 1 2 3 4 5 6 Delivery Location 0.438 (4.529, 0.000) 0.414 (4.203, 0.00) 0.372 (3.747, 0.000) 0.351 (3.464, 0.001) - - Toilet Access - 0.148 (1.284, 0.199) 0.005182 (0.041, 0.967) -0.009443 (-0.074, 0.941) - - Roof - - 0.288 (2.708, 0.007) 0.269 (2.501, 0.013) 0.359 (3.786, 0.000) - Educational Attainment - - - 0.115 (1.116, 0.265) - 0.267 (2.733,0.006) N 798 798 798 798 801 805 Adjusted R2 0.024 0.025 0.032 0.033 0.018 0.009

From this summary table it is clear that delivery location is significantly associated with waz and remains significant after controlling for sanitation, SES and educational attainment. As the delivery location coefficient changes little between models, there is no confounding between the other independent variables and delivery location. However, if you look at toilet access in models one and two, the coefficient changes by a large degree, indicating confounding, but note that it is  insignificant in each model.