Yes, conductive media usually have a measurable temperature co-efficient.
However, it is possible to externally control the temperature of the resistive element, and at low power the self-heating should be negligible.
As I pointed out near the beginning of this thread, Ohm's Law is an approximation that is usually useful, but it is not strictly exact: many resistors exhibit a voltage co-efficient of resistance (unrelated to temperature), but it is rarely specified on the data sheet.
I have seen physically small (0603 and 0805 SMT) high-value resistors where the resistance changes by > 0.1% at constant temperature for only 10 V applied to 50 megohms.
Lower-value resistors in reasonable packages (e.g., RN60 metal film TH resistors) are more linear.
Light bulbs are extremely non-linear, treated as components, over their usual operating voltage and current range, as shown in measurements reported above.