# Why is sample selection based on the dependent variable problematic?

Why is sample selection based on the dependent variable problematic? Why is sample selection based on the independent variable not problematic? Consider the following function. $$saving=\beta_0+\beta_1income+\beta_2age+\beta_3size+u$$ If our data set was based on a survey of people over 35 years of age, thereby leaving us with a nonrandom sample of all adults, the coefficient estimators are still unbiased and consistent. Consider another function $$wealth=\beta_0+\beta_1educ+\beta_2exper+\beta_3age+u$$. Suppose that only people with wealth below \$250,000 are included in the sample. This will result in biased and inconsistent estimators of the parameters. Why is that the case? Can you explain intuitively?