A crude way to verify tempco is to heat up a resistor while measuring it, for example with a hot air rework station. Set the temperature to for example 70°C. This heats up the resistor by about 50°C. A good metal film resistor would have about a 0.1% increase in resistance. Crappy carbon film might have a decrease of about 1%. Even if you don't have an accurate way to control the temperature, you can just connect several resistors (eg. known good metal film, known carbon film, unknown) to separate multimeters (one ought to be enough for anyone, right?), and heat them up simultaneously. There should be a big difference between the metal film and carbon film.
I'm not sure how much of an issue this is these days. A few years ago several Chinese vendors, including Sure electronics, were selling carbon film resistors as metal film.