Analysis for two continuous variables (not normally distributed)

[Yue YY]

Dear Statlist,

I have two continuous variables, the dependent variable is not normally distributed while the independent variable is normally distributed. Apart from Spearman's correlation (I actually am not sure if this is suitable since the dependent variable is not ordinal), I'm wondering which method I can use for analyzing the relationship between this two variables? Thank you very much.

Yue

[Marcos Almeida]

There is no need for the DV to be normally distributed so as to perform a regression analysis.

[Yue YY]

There is no need for the DV to be normally distributed so as to perform a regression analysis.

Thank you Marcos! May I ask one more question: is that necessary to have DV normally distributed for multiple linear regression?

[Carlo Lazzaro]

Yue:
no, it is not (as Marcos clearly highlighted).
In both simple and multiple OLS, normality (and omoskedasticity) relates to residual distribution only.

[Yue YY]

Yue:
no, it is not (as Marcos clearly highlighted).
In both simple and multiple OLS, normality (and omoskedasticity) relates to residual distribution only.

Thank you very much, Carlo!