[magnusbaringer]

Thank you very much for your time and support!

I do understand your concern - however the data collection works a little bit different "under the surface" as it seems. In the statistical analysis of the data we do not look at the "complete" breast images. The study is designed in a way, that the most important parameters around breast (reconstruction) surgery (e.g. breast size, breast shape, areola size, areola position, scars etc.) are being combined in 88 images in a way that in the end allows us to extract the data "broken down" to the different parameters. So the result of the study will not be "people like image X more than image Y". Instead we want to get a far more profound understanding of the influence the different parameters have on the perception of the breast. As we also measure the decision time, results of the study will be something like: "most people (or maybe just men - that's what we want to find out) tend to accept breast size asymmetry more then areola asymmetry. The decision happens much faster on areola asymmetries - so that seems to be more obvious and therefore has to be taken more care of during the surgery." Nobody has every done a profound analysis of these individual parameters alone - so that is where we see huge potential for a better understanding. I hope that explanation helps a little bit! Thank you very much again for your support!

[dotancohen]

  > As we also measure the decision time,

From very much experience in designing and running online interactions, including polls, I can tell you with authority that you cannot trust the decision time of answers online. You don't know that they user didn't look at a quick text message on his phone, or adjust the music volume, or go to the bathroom, or take a call, or had a network issue, or your images didn't load quickly, or the dog didn't bark suspiciously, or a loud C-130 just flew overhead, or any other of a million things. I highly suggest disregarding that feature.

[pinko]

Isn't that just noise? It's possible to extract useful signals from extremely noisy data if you understand the statistics.