True or false and why? a) With Panel data, Least Squares Dummy Variables fixed effects estimation and within- transformation give exactly the same coefficient estimates for explanatory variables and the same R-squared statistic. b) It is possible for the Logit model to give a predicted likelihood of an outcome, that is larger or equal to one, or 100% or more. c) Suppose that you have estimated P(Y) = G[0.4 - 1.2X] using a Logit model, where G is a cumulative logistic function, G(z)=exp(z)/[1+exp(z)], and that the sample mean of X is 0.56. The predicted likelihood of Y happening when X is 0.5 is .450166. d) Instrumental Variables estimation should not be used if the Instruments are not normally distributed, because the resulting t-statistics will be wrong. e) The Instrumental Variables estimator can be biased. f) To compute the marginal effects of the estimates in a Probit model, one needs to make some assumptions about what values the explanatory variables take. g) In a Randomised Controlled Trial (RCT), "attrition' means that some of the observations are lost from the sample in during the trial. h) In difference-in-differences analysis, the main assumption for identification is that the average pre-treatment characteristics of the treatment and control groups are similar