Our null hypothesis is that the true coefficient is zero. In which case would we choose to use a one-sided alternative hypothesis over a two-sided alternative hypothesis? We choose a one sided alternative hypothesis when we have a large sample size. We choose a one sided alternative hypothesis when we specify a low (0.01) significance level. It does not matter what type of alternative hypothesis we specify. Choose a one sided alternative hypothesis when the theory and literature suggests either that the true coefficient should be greater than zero or that the true coefficient should be less than zero.

In the p-value approach to hypothesis testing, we reject the null hypothesis if O p-value > 0.05 O p-value 0.05 O p-value 0.05 O p-value < 0.05

Suppose we have sample data on (y,x) of size n. Consider regressing y on x and a constant using the least square method (OLS estimation). The estimated equation is y = 2.59 +5.19 x + e where e is the residual. The standard errors for the constant and the coefficient were 0.51 and 1.58 respectively. To test the null hypothesis of zero coefficient (for x) at a given significance level, we may use t-statistic. Compute the t-statistic for the coefficient. Make sure the answer is round up to the nearest hundredth. e.g. 3.015 --> 3.02

Spatial Resolution: Regionally specific fine-scale features and changes are difficult to capture in global climate models since they usually operate at coarse spatial resolutions. The inability to produce precise and comprehensive forecasts for localised consequences is a result of this restriction. Generate a hypothesis related to this problem