Multiple Regression Models

Papers often report several regression models next to each other. This table shows a few truncated models from the example in the video, about the relationship between a country's property rights protections and its GDP. Model 1 shows the coefficient and standard error for the effect of property rights protection on GDP without any control variables. Model 2 shows the effect of property rights protection on GDP while controlling for the country's latitude, and Model 3 shows the same, while also controlling for whether the country is in Asia. Why might the effect of property rights protections on GDP decrease as we add more variables to our model?

Variable	Model 1	Model 2	Model 3
Protection	.52(.06)	.47(.06)	.43(.05)
Latitude		.89(.49)	.37(.51)
Asia Dummy			-.62(.19)

Possible answers

Because the inclusion of more variables weakens the statistical power of an independent variable

Because omitting these variables caused the effect of property rights protection on GDP to be downwardly biased.

Because the inclusion of more variables strengthens the statistical power of an independent variable

Because omitting these variables caused the effect of property rights protection on GDP to be upwardly biased.

Exercise

Multiple Regression Models

Instructions

Possible answers