1. Assume that you have been researching an issue. You have generated 1290 obser
ID: 1114828 • Letter: 1
Question
1. Assume that you have been researching an issue. You have generated 1290 observations with 4 explanatory variables. When you run your regression, the estimated coefficient for one of your explanatory variables has the wrong sign. Although the regression indicates that the explanatory variable is inversely related to the dependent variable, you know that the true relationship between the two variables is positive What might have caused the problem? Is there anything that you can do to correct the problem? If so, what should you do?.lf you already know that the relationship is positive, why include it in the regression?Explanation / Answer
The problem of unexpected signs may be due to the problem of stationarity which leads to the problem of unexpected signs. To get rid of this problem one must use differenced -variable approach to regress the model.
Another problem that could lead to this issue may be the endogeneity problem where x's are related to error term. one should use the instrumental variable approach to fix this.
Related Questions
Navigate
Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.