Unfortunately I don't see being able to just scale things with simple resistors as particularly safe. You can get close but it would require a pull-up resistor to a voltage >> the 5V the MA3 is operating on which could be painful to the MA3 if something wasn't connected correctly or came loose. Also sort of an unknown is the output impedance of the MA3 so it's difficult to calculate the exact resistors needed since that would factor in.
An example would be to use a regulated 7.2V reference with a voltage divider set between the 7.2V and the MA3 output consisting of a 10K and 13K resistor.
7.2V --- 10K --- x --- 13K --- MA3 output, the 'x' junction between the two resistors being your output to the radio.
At 0V input the voltage divider sees 7.2V so (using 10K at the top and 13K at the bottom) the input to the radio sees 7.2V * 13K / (13K + 10K) = 4.07V.
At 5V input the voltage divider sees (7.2-5.0) = 2.2V so the radio sees the 5V + (2.2V * 13K / (13K + 10K)) = 6.23V.
It's pretty close and if you used a small potentiometer at the 'x' you could probably tune it in.
You also need to make the 5V for the MA3 and the 7.2V reference voltage... depending on what you have for a battery... if you have a 9V battery you could probably use LM317 adjustable regualtors to make the voltages easily enough.
I guess a reasonable question is just how precise you want this to be. I worked out a circuit using op-amps and a single 9V battery supply but it's not trivial to build if you've never wired up a breadboard/plugboard before. That circuit actually uses 5V as a reference so you kill two birds with one stone since that same 5V could power the MA3 easily enough.
With respect to 'eating'... I'm at work guys. Some days are more involved than others.
