Most of the temperature drop is from fuel evaporation downstream of the venturi, causing ice from ambient humidity to form on the throttle plate (injected engines also have venturis and throttle plates for fuel and air metering, but because no fuel is being introduced there, so there is no evaporative cooling, there is no need for carb heat and they just use an alternate air source in case of main intake obstruction from other causes).
The temperature drop itself varies with the amount of fuel being introduced, humidity, and throttle opening and like most natural phenomena like this, it forms a bell curve of temperature/humidity, with the middle of the bell curve (highest risk of carb ice) at around 15-20 deg C. Problem is, that's the middle of the bell curve and the edge of the bell curve can extend to 30 deg C and beyond and you just don't know where you are in it. On top of that, if it's 30C on the ground, it's going to be just over 20C at 4000 ft, getting you closer to the middle of the bell curve.
So it's somewhat less likely when it's that hot outside, but can still happen and unless you have a carb air temperature gauge that you trust, this is something you don't screw around with and you follow the normal carb heat protocol regardless.
There is a similar situation with the engine itself. Lycomings are somewhat less prone to carb ice than Continentals because of the way the carburetor connects to the engine (the Lyc carb gets more conducted and radiated heat from the engine because the intake spider runs through the sump, so its body is slightly warmer than a Cont's), but they can still get ice and you follow the same carb heat procedures on either engine regardless.