Is it possible for an F-statistic to be negative? Justify your answer. (Hint: In answering your question, assume that the number of observations is larger than the number of parameters we are trying to estimate. This will be true of any reasonable regression analysis to avoid overfitting.)