Wait, my advice does not work in such a case, because $(\psi(-z)+\gamma)^2$ has a double pole in $z=n$, due to:
$$ g(z)=\gamma+\psi(-z)= \frac{1}{z}+\sum_{n\geq 1}\left(\frac{1}{n}+\frac{1}{z-n}\right).$$
Anyway, given that $z=n$ is a double pole for $(\psi(-z)+\gamma)^2$ and $h(z)=\frac{1}{(z+1)(z+2)^3}$ is holomorphic over $\Re(z)>-1$, then $z=n$ is a double pole for $f(z)$, so:
$$\operatorname{Res}\left(f(z),z=n\right) = \frac{d}{dz}\left.(z-n)^2 f(z)\right|_{z=n} = \left.2(z-n)f(z)\right|_{z=n}+\left.(z-n)^2 f'(z)\right|_{z=n}\tag{1}$$
where:
$$ \left. 2(z-n)f(z)\right|_{z=n} = 2\operatorname{Res}(f(z),z=n)=\frac{2}{(n+1)(n+2)^3}$$
by the correct version of my advice (simple poles), and:
$$ \left.(z-n)^2 f'(z)\right|_{z=n} = \left. (z-n)^2\left(g'(z) h(z)+g(z) h'(z)\right)\right|_{z=n}=h(n)\cdot\left.(z-n)^2g'(z)\right|_{z=n}.\tag{2} $$
Since:
$$ g'(z) = -\sum_{n\geq 0}\frac{1}{(z-n)^2} $$
we have that $(2)$ just equals $\frac{1}{(n+1)(n+2)^3}$, so:
$$ \operatorname{Res}\left( f(z), z=n \right) = \color{red}{\frac{1}{(n+1)(n+2)^3}} \tag{3} $$
like in the case when $g(z)$ was not squared.
$2\mathrm{Res}(f(z), z =n)$$ Do you mean the SINGLE ORDER residue? How? The function $f(z)$ is still of order two poles.?
– Amad27 Jan 15 '15 at 14:05