x
We use cookies to create the best experience for you. Keep on browsing if you are OK with that, or find out how to manage cookies.

# Applied econometrics Essay Example

Show related essays

## Applied econometrics

Thus, lnchpct accounts for the most change in model (3), more specifically 63. With the change (decline) in the coefficient estimate for studpc, it shows that the effect of studpc declines as more variables are controlled for its effect on totsc4 from model (1) to model (3). The additional variables improve the model by explaining much of the variance in totsc4 and thus the estimate is closer to the truth in model (3). An improvement in the R2 measure does not guarantee the use of logarithmic form of totday without looking at the model fitness. What should inform the use of either level form or logarithmic form of the variable is whether by using level or logarithmic form, the model fitness improves or not. Thus, if the R2 increases and the model is also fit, then we can use the logarithmic form of the variable. If the new model does not fit, we cannot use the logarithmic form of the variable. As explained in Q.7 above, the new model would only be said to be better if the fit is improved, not necessarily the R2, which has only insignificantly improved from 2.1% to 4. If the model fit improved in the process of using the logarithmic form as well as the level form of totday, then it can be said that the model is now better. The two models are now incomparable as one is in a logarithmic dependent variable form and the other in the level form. To make a decision, the R2 has to be adjusted in the equation with the logarithmic form of the dependent variable. First, we have to obtain the fitted values of ln_totsc4 from the regression results.These fitted values are then used to compute an estimated average value for the dependent variable by taking antilogs and making the bias adjustment. Then, the square of the correlation is calculated between the dependent variable value and the fitted values for the dependent variable.This is now directly comparable to the R2 of the original model. Another method that can be used is to judge the models by their ability to predict the dependent variable. Here, the competing MSE or RMSE are calculated and the model selected is the one with the lower forecast error. Ramsey’s RESET test for a model tests the hypothesis the null hypothesis that the regression is correctly specified. Thus, the null hypothesis is usually rejected when the p-value is less than the desired level of significance. In this case, the p-value is

Close ✕