If the integral looks like this: $$ \iint \{\iint f(x)f(y)\mathrm{d}x\mathrm{d}y\}g(x)g(y) \mathrm{d}x\mathrm{d}y $$ is it legitimate to 'combine' the integrals as $$ \iint f(x)f(y)g(x)g(y) \mathrm{d}x \mathrm{d}y ? $$
The assumption is that both $f$ and $g$ are non-negative, and are elements of $L^2([0,1])$ space ( i.e. they are square-integrable).
Of course not. Just take $[0,1]$ and $f=0$ on $[0,1/2]$ , $f=1$ on $[1/2,1]$. Now take $g=1-f$. Then $\int\int f(x)f(y)g(x)g(y)dxdy=0$, whereas the quadruple integral is not zero.