Still, he says, "I have some concerns about whether or not the particular folktale dataset he has used is appropriate for phylogenetic analysis".
For the van de Vijver dataset, he outcome variable (survival) is right censored, i.e., if a patient survives for > =5 years, she is considered to have good prognosis, else bad prognosis.
"Why not leverage Tinder to build a better, larger facial dataset?" he asked.
Ross Kaminsky, a columnist on American Spectator, claimed: "One implication, supported by Briffa's near decade-long refusal to share his data, is that he cherry-picked the dataset that supported the conclusion he wanted to find".
He hasn't made the dataset available - perhaps because it's far too valuable to a venture capitalist like himself.
The dataset was scored by a single rater (HE).
After one collects information from the user, he or she can save the dataset.
"I'm going to make it my mission to talk to English Heritage to reveal the dataset in a way that's useful to people," he says.
For every individual recruited on MTurk, he harvested information about their friends, meaning the dataset ballooned significantly in size.
He also observed that, "Beyond this, there is clear noise in the dataset.
