I have to find a explicit form of probability distribution of $X^3$, if $X \ \mathtt{\sim} \ U[a, b], \ -\infty < a < b < \infty$.
So far I've succesfully done a simpler version, when it's $\ U[0, a], \ \ 0 < a < \infty$:
$F_{X^3}(x) = \mathbb{P}(X^3 \leq x) = \mathbb{P}(X \in [0, \sqrt[3]{x}]) = \int^{\sqrt[3]{x}}_0 \frac{dx}{a} = \frac{\sqrt[3]{x}}{a},$
$f_{X^3}(x) = \frac{dF_{X^3}(x)}{dx} = \dfrac{1}{3ax^{\frac{2}{3}}}$, where the support is $0 \leq x \leq a^3$.
However, the way of thinking that is above got me lost at the full version of the uniform distribution.
$F_{X^3}(x) = \mathbb{P}(X^3 \leq x) = \mathbb{P}(X \in [-\sqrt[3]{x}, \sqrt[3]{x}]) = \int^{\sqrt[3]{x}}_{-\sqrt[3]{x}} \frac{dx}{b-a} = \frac{2\sqrt[3]{x}}{b-a}$
$f_{X^3}(x) = \frac{dF_{X^3}(x)}{dx} = \dfrac{2}{3x^{\frac{2}{3}} \cdot (b-a)}$, where the support is...?
If the support would be $a^3 \leq x \leq b^3$ and $a < 0$ then the integral $\int^{b^3}_{a^3} f(x)$ would have no sense for negative values, whereas it should equal $1$. Or where else did I made a mistake?
If we want $X^3 \leq x$ then we are asking that $X \in [-\infty, x^{1/3}]$. Support of $X$ is $[a,b]$ so we can consider any $x$ so long as $a^3 \leq x \leq b^3$.
Hence the integral limits for $F(x)$ are $a^{1,3}$ and $x^{1/3}$.