A question on effect on the dependent variable, coefficient estimates, concept of omitted variable bias. see below please

Consider the paper by Achen and Bartels (2012) that relates shark attacks to voting patterns amongst New Jersey counties at the time of the 1916 US presidential elections. Votei = a + beachi + machinei + OWilsoni + Ei where Votei is the county-level vote share for Wilson in 1916, Beachi is an indicator variable for whether a county contains a beach, Machinei is another indicator variable representing counties where individuals hostile to Wilson are in control of the Democratic Party, Wilsoni is the vote share for Wilson in the previous 1912 election; and Ei is an error term. The main variable of interest is beachi since the authors' hypothesis is that counties represented by this variable are most sensitive to concerns about local shark attacks. (a) (c) Consider the following results from estimating the baseline model outlined above: -0.032 (0.010) for and -0.056 (0.011) for , where the number in brackets represents the standard errors. How do we interpret the estimated coefficients for and in terms of some potential effect on the dependent variable? (5 marks) We could run the baseline specification again with 'robust' standard errors. What econometric problem would this assist us with? And would this change to our empirical model affect both the coefficient estimates and the standard errors? Explain. (5 marks) Suppose that we include an outlier county that was initially excluded from the baseline model. How could this change affect either the coefficient estimates or the standard errors? (5 marks)