Lab 4

Which of the following is not an assumption of linear regression?

• Normal distribution of errors

• No perfect multicollinearity

• Equal variance in groups

• Homoscedasticity

This distribution fulfills the assumption of homoscedasticity.

• True

• False

In a model, we include both the regional and country characteristics as independent variables. Which of the following is true?

• We do not have to check for linearity.

• This could lead to collinearity issues.

• Outliers are controlled for.

• We have an iid sample.

If linearity does not hold...

• one should not analyze the data at all.

• one should always use robust standard errors.

• one should transform the variable accordingly.

• this is not problematic.

Which of the following statements is true?

• cons_ is the predicted outcome if treatment = 0

• the estimate of population is not significant.

• treatment has the smallest effect on the outcome.

• 38% of variance in turnout are explained by this model.

An interaction effect is a way to formalize a moderating effect of one variable on the relationship between two variables.

• True

• False