I had to think about that for a while but I see you are right. So my complaint instead is that my 'true RMS' multimeters read 1.5V for the same waveform (on the DC volts setting). This bothers me. I suppose they are discarding the AC component of the waveform before measuring, although that's not really what I would expect from a notionally DC-RMS voltmeter measuring a unipolar waveform!
Well, the
average DC voltage on that 0V-to-3V signal is 1.5V indeed. You can interpret it as an AC-only signal (with +- 1.5Vpp) on top of a 1.5V DC offset. For a multimeter, I would indeed expect it to give me the DC offset (i.e. average DC voltage) when I measure the signal in DC mode, and the RMS voltage of the AC-only component when I measure in AC mode. So your multimeter probably performs as it should.
The scope gives you more flexibility, as you can choose to measure either the
average voltage
or the RMS voltage of the "mixed" 0V-3V signal. The multimeter does not have these two separate settings, I assume, so it opts for the more commonly required readout, namely the average DC.