I don't know about it being worse than Mac or Linux but yes, using the volume control or the Windows stack will measurably degrade performance. I plan to show that in a future article.
Briefly, when you use the audio channel in Windows normally, all sound gets converted to floating point internal representation. They are also converted to a single sample rate and bit depth defined in the Sound control panel for your default output device (same thing happens in MacOS). Mixing of sound from multiple apps and volume control is performed. Then the final data is converted back to PCM integer data with dither and sent out to the sound card. All of this processing will change the performance of the system and hence the reason people will say this pipeline is not "bit exact."
WASAPI and ASIO interfaces were designed to get around this issue. When you use them in exclusive mode, you will get a direct pipeline to the sound card and no processing occurs. You will lose Windows volume control and any mixing of sound from multiple channels. But you get "bit exact" performance. For best fidelity this is the mode you should be using and is how I test everything.
MacOS/iTunes do similar thing unless something has changed in recent releases.
BTW, the above is our new member, j_j's doing so he can explain more.