wolfadeus - 8:14 am on Aug 14, 2012 (gmt 0)

Being a fan of empirically gathered evidence, I recommend you the following as my final reply: Use CSS to test three or four different colour schemes on your specific target groups. You can randomise the colours for each user and then measure the conversions your reach. You might have to run this for a few weeks, but by the end, you will know if there is a significant difference.