Sorry for the confusion! What I meant by the first part was that if you're given a question like ∫ (-3/2x^2)dx, you can simplify it first by dragging -3/2 outside of the integral sign which gives you -3/2 ∫ 1/x^2 , and then you bring the denominator of whatever's inside the integral, up to make it an exponential (and then you can do whatever you need to, to solve the question
). Whereas with something like ∫(5/3x)dx, you don't bring the 5/3 outside of the integral, but instead, you look at it in terms of logs.
My issue is that I can never tell which way I should approach the question, I'm actually terrible with maths 
The question's mostly addressed above but I'm gonna make a remark. The main reason why that method works for the question you wrote down is essentially because it was of the form \( \int x^n\,dx \) where \(n \neq -1\). In your case, you had \( n = -2\).
So by quoting that rule, \( \int \frac{1}{x^2}\,dx = \int x^{-2}\,dx = -x^{-1} + C = -\frac1x + C\).
Whereas when \( n = -1\), we have \(\int x^{-1}\,dx = \int \frac1x \, dx = \ln(x) + C\). In general, when the derivative of what's on the bottom appears in the top, you have those integrals that go to a log. But for the other one there was no problem isolating a power on \(x\)
all by itself, so you should be using the power law instead.
Note also (from the reference sheet) that the same thing can be done with integrals of the form \( \int (ax+b)^n\,dx\), for example, \( \int \frac{1}{(9x-1)^2}\,dx\).