Worth noting that this study, which I just happened to read today, is small (n=234, with the number of useful responses even lower). And the ‘up to 20%’ refers

Worth noting that this study, which I just happened to read today, is small (n=234, with the number of useful responses even lower). And the ‘up to 20%’ refers to the fact that they had some odd responses which indicated that the time wasted was the entire hour of the study. If they corrected these down the figure becomes 11%. Oh and it’s self reported. From participants crowdsourced at Prolific, who were paid. It didn’t seem like the most rigorous study to me. And I’m invested in the result being bad, as the thing I do claims to solve some of these problems!