A Poisson regression model is fit, relating total goals scored per game among 5 premier football leagues: England Cormany, Spain, Mexico, and Brazil. 1752 games are observed across leagues for the 2013 seasons. The model uses England as the "reference" country and has dummy variables Xo, Xs, XM, and XB, respectively. That is: = 1 if Germany, O if not As = 1 if Spain, 0 if not XM = 1 if Mexico, 0 if not X, =1 if Brazil, 0 if not differences) The link function is given below for the null model (no country effects) and the complete (saturated) model (country Null Model (modo): log ( /) = Bo Saturated Model (modl): log( H) = But BoXe+BsXs+ BMXm+ BBXB The following results are obtained using R. > modo summary (modo) Coefficients : Estimate Std. Error z value Pr(> |z|) (Intercept) 1.00096 0. 01448 69.11 modi summary (mod1) Coefficients : Estimate Std. Error z value Pr(>|z|) (Intercept) 1. 018277 0. 030831 33.027 XG 0. 132336 0. 044550 2.971 0. 00297 * * XS -0. 006676 0. 043675 -0.153 0. 87851 XM -0. 099375 0. 047480 -2.093 0.03635 * X B -0. 116833 0. 044933 -2. 600 0. 00932 Null deviance: 2113.5 on 1751 degrees of freedom Residual deviance: 2077.5 on 1747 degrees of freedom > logLik (mod1) 'log Lik. ' -3333.973 (df=5) p.2.a. Compute the predicted total number of goals for Mexico games based on each model (in the actual goals scale). Saturated Model: Null Model: p.2.b. Test whether the true population means are equal among all leagues