Here is a set of nested models:

Dependent Variable: Respondent's Income in 1000's

Independent Variable: Model 1 Model 2 Education (in years) 3.57*** -3.57*** Education Squared --- 0.28*** Hours Worked 0.71*** 0.71*** Constant -38.60 2.55 R-Squared 0.20 0.30

What is the most appropriate conclusion to make about these models?

Because the linear effect does not lose its statistical significance in Model 2, the linear model is the better model.

Neither the linear model nor the non-linear model is better.

Because the squared slope is statistically significant, the non-linear model is the better model.

Because the R-squared for the non-linear model is much higher than the R-squared for the linear model, the non-linear model is the better model.