I don't know if the sounds on the article are the same sounds they used in the study, but their example of a consonant chord doesn't sound particularly consonan

I don't know if the sounds on the article are the same sounds they used in the study, but their example of a consonant chord doesn't sound particularly consonant to me. (The dissonant chord sounds like it has two notes close to each other, which is almost universally objectionable.) I'm not sure why not; it could be the intervals are tuned in some weird way, or for some reason they chose a timbre whose harmonics aren't in a whole-number-ratio with the fundamental. This is the sort of thing where the experimenter has to be careful about what they're testing. To someone unfamiliar with 12-tone equal temperament, a major chord on a piano might sound out of tune relative to just-tuned 4:5:6 major chord. A just chord can also sound out-of-tune when played in an inharmonic timbre. If you play a major chord on a piano and a listener says "I don't like how that sounds", it might not be that they don't like harmonic music. It might be just that they don't like how a major chord sounds when played on an equal-tempered piano. One way to do this kind of testing is to use pure sine waves that have no harmonics. So, without harmonics there's no particular reason to prefer just intervals to random, arbitrary intervals except that the former may conform to the listeners expectations.