On 2017 Apr 17, Lydia Maniatis commented:

To synopsize the problem as I understand it: The author is claiming that large sections of the population are walking around, in effect, with differently tinted lenses, some more bluish, some more white or yellowish.

This is a radical claim, and would have wide-ranging consequences, not least of which is that, as in the case of the dress, there would be a general disagreement about color. Such a general disagreement would not be detectable, as no one could know that what we all, for example, might call blue is actually perceived in different ways.

The reason we can know that we disagree about the colors of the dress is that we agree on colors generally, and the dress constitutes a surprising exception.

If the author believes in his hypothesis, a strong, direct experimental test is in order. (It would certainly falsify.) If he insists on focussing on correlations with "owls" and "larks," then he should better control his populations, e.g. use night watchmen for the owls, and park rangers for the larks, or investigate how the dress is perceived by populations in e.g. Norway, where the days and nights are months-long and the same for everyone. Do we get less variation there?

What doesn't seem worth pursuing is another uninterpretable replication based on poor quality, muddy and uncheckable data from anonymous readers of Slate.

On 2017 Apr 15, Lydia Maniatis commented:

Please see comments/author responses on PubPeer.

On 2017 Apr 15, Lydia Maniatis commented:

Please see comments/author responses on PubPeer.

On 2017 Apr 17, Lydia Maniatis commented:

To synopsize the problem as I understand it: The author is claiming that large sections of the population are walking around, in effect, with differently tinted lenses, some more bluish, some more white or yellowish.

This is a radical claim, and would have wide-ranging consequences, not least of which is that, as in the case of the dress, there would be a general disagreement about color. Such a general disagreement would not be detectable, as no one could know that what we all, for example, might call blue is actually perceived in different ways.

The reason we can know that we disagree about the colors of the dress is that we agree on colors generally, and the dress constitutes a surprising exception.

If the author believes in his hypothesis, a strong, direct experimental test is in order. (It would certainly falsify.) If he insists on focussing on correlations with "owls" and "larks," then he should better control his populations, e.g. use night watchmen for the owls, and park rangers for the larks, or investigate how the dress is perceived by populations in e.g. Norway, where the days and nights are months-long and the same for everyone. Do we get less variation there?

What doesn't seem worth pursuing is another uninterpretable replication based on poor quality, muddy and uncheckable data from anonymous readers of Slate.