Which of the following are true?

Removing explanatory variables from a regression will...

i) never make coefficient estimates more significant

ii) always make coefficient estimates more significant

iii) change the interpretation of our model coefficients

iv) always make the standard errors decrease

v) sometimes lead to overfitting the model, leading to poor out of sample predictions

vi) never increase the R^2

vii) always make the standard errors increase

viii) sometimes change the coefficient estimates on existing explanatory variables