To be even more pedantic, what does f^{-1} actually mean in your context? It's a very annoying notation that could mean two different things, but it's historic, so we don't care to fix it. I'll take that you mean "the inverse function of x"
As of this hour, if f(x)=x^2-5/4 (and supposedly defined over all real numbers), the inverse function f^{-1} doesn't even make sense, as davros stated. We need some domain on which f is defined, so that f is bijective. If you don't know what bijective means, pretend it's "strictly increasing/decreasing in the domain" for now. The key point is the inverse function, at the end of the day, is still a function - for every number from the domain, the function outputs one and only one number. If it outputs more than one number or no number, it's not a well-defined function to begin with.
As a sidenote, equally as annoying is "sin^{-1}(0)". This, we take sin^{-1} to be "arcsin", which actually outputs a specific range of values so that it makes sense (i.e. your calculator is correct). But as we know, sin(0)=sin(180)=sin(360)=sin(-180)=...=0.
-- downstairs is not even a thing in GCSE/AL, I think --
But if they mean f^{-1}({19}), now it becomes the pre-image of f of the singleton set {19}. This notation now makes sense, as it really means "what input does f outputs 19", and the answer is indeed the set {4.5, -4.5} (not "4.5 or -4.5", the answer should be a set).