Thank you Dick Pierce for your thoughtful reply to my question. Being a
"digital" power meter, I have assumed that the "averaging" process is
carried out digitally since we are talking of integration times up to 200
seconds.

Presumably, the meter takes the pressure from its microphone, converts that
linearly to voltage, then performs the operations of squaring the voltage to
get the power, then compressing this power logarithmically to arrive at dBs.
Just where in this sequence, the "average" is taken is the point on which I
am not clear.

More careful experiments, where I took "averages" over times up to 120
seconds while turning the noise source on and off with a 50% duty cycle,
yielded an "average" reading that was approximately 6 dB below the noise
source power. This would suggest to me that the averaging is being done on
the microphone output voltage (half voltage equals -6db power drop).

However, It would be great if someone has actual knowledge of the power
meter circuitry and its functions. Any offers?

Hank