This comes from here:
The NPR story:
Reading the above, one might think they actually ran a study where they replaced/substituted foods. But they didn’t. Instead, they took studies (epidemiological studies based on FFQs, food-frequency questionnaires) and ran them through a model that is supposed to estimate what happens if you substituted foods.
They did the same with "green house gases’, likely with a heavy bias towards indicating that plant-based is somehow better for the planet than is meat-based.
But I can guarantee that if you actually had a randomized controlled trial where people actually ate more red meat as compared to people who drank soy based drinks, you wouldn’t get these results. This assumes you made the study long enough and you kept the red meat part low carb (if both groups eat high carb but one eats some red meat and one drinks soymilk, and you look for short term metrics like LDL, the study might actually favor plants).
And as for the effects of these on green-house gases, that is so complex, I’m not even going to comment.