Which of the following are true?

Multivariate linear regression...

i) is always better than univariate regression for answering causal questions because including more variables leads to a higher R^2

ii) is always used to do prediction, and cannot be used for assessing the impact of different treatments in an experiment

iii) is not susceptible to biased coefficients due to measurement error

iv) gives parameter estimates that are proportional to partial correlations between y and each of the x variables

v) can be used to determine average treatment effects with several arms of treatment

vi) doesn't suffer from omitted variables bias if we interact all the explanatory variables currently in the model

vii) let us use dummy variables to model different intercepts and slopes for different categories in the data