Based on Analysis C and D, we can conclude other risk factors are NOT potential confounders for the effect of BMI and ethnicity on estradiol level according to the 10% rule (i.e. the regression coefficient estimate changes >10% vs. F 0.0090 Residual 151360.426 207 731. 209787 R-squared 0. 1052 Adj R-squared 0. 0923 11 1 Total 169162. 084 210 805.533732 Root MSE 27.041 Estradl Coef. Std. Err. t P>It| [95% Conf. Interval] Ethnic -81. 24498 24.51533 -3.31 0.001 -129.5767 -32. 91325 bmi -2.388122 .9406869 -2.54 0. 012 -4.242677 - . 5335669 intethnicbmi 2.720802 1. 019331 2.67 0.008 . 7112016 4.730403 _cons 110. 3196 22. 02682 5.01 0.000 66.8939 153.7452 Analysis D . reg Estradl Ethnic bmi intethnicbmi Entage Numchild Agefbo Anykids Agemenar if Numchild !=9 & Agefbo !=99 & Anykids !=9 & Agemenar !=99 Source SS df MS Number of obs 203 11 F(8, 194) 3.67 Model 22073.8037 8 2759. 22547 Prob > F 1I 0.0005 Residual 145860.936 194 751. 860496 R- squared 0. 1314 Adj R-squared II 0. 0956 Total 167934.74 202 831.360099 Root MSE 27 .42 Estradl Coef. Std. Err. t P>It| [95% Conf. Interval] Ethnic -80.90427 25. 36418 -3.19 0.002 -130.9292 -30. 87932 bmi -2.564426 .9679979 - 2. 65 0. 009 -4.473577 - . 6552749 intethnicomi 2.727336 1.051687 2.59 0. 010 . 6531275 4.801544 Entage . 8011156 . 4055913 1.98 0.050 . 001181 1. 60105 Numchild -1. 135726 3.496057 -0. 32 0. 746 -8. 030886 5.759435 Agefbo -. 9429103 . 8915768 -1.06 0. 292 - 2. 701338 . 8155177 Anykids 19.30578 22.8085 0. 85 0. 398 -25. 67868 64. 29024 Agemenar . 5246202 1.453864 0. 36 0. 719 -2. 342789 3. 39203 _cons 88. 10261 29.96457 2.94 0. 004 29. 00446 147 .2008