For reference, I was in a PhD program in the humanities at the University of Wisconsin, and it was definitely not fully funded.

In STEM fields, 10% of PhD students reported using external (non-university) funds as their primary source of grad school funding; 10% of those externally funded students (1% overall) reported using loans as their primary funding source.

In non-STEM fields, the proportion of non-funded PhD students is higher, at 35%, of which about half (17% overall) used loans as their primary source of funding.

Given that the majority of PhDs in the US are in STEM, it is correct to say that the vast majority of American PhD programs are fully funded.

No. The link is a study of PhD recipients not PhD students. This is a crucial distinction, because my main claim was about the effects of attending graduate school on non-recipients. It's quite likely that PhD recipients are better funded than non-recipients, which funding helps them to complete their PhDs. In my academic department, there were many more students than there were sources of funding such as teaching assistantships, research assistantships, lectureships, and fellowships.

> In non-STEM

It's not clear to me that the study was comprehensive. It says STEM and SBE (social, behavioral, and economic), but does SBE mean non-STEM? Why didn't they just say non-STEM? Does SBE include Art, English, History, and Music, for example?

> Given that the majority of PhDs in the US are in STEM

~65% of PhDs awarded are in STEM, but only ~25% of Masters degrees are in STEM. This may say a lot about the PhD completion rate. At my school, a Masters was often rewarded on the way to a PhD, so it was kind of a (mostly worthless) consolation prize for those who didn't finish their PhD.

Anyway, the study is primarily about the "primary source" of funding, which is not quite the same as full funding.

The linked study was heavy with survivorship bias. Sadly, grad school dropouts are mostly forgotten and ignored by society.