I think those points are all valid. Additionally I think the fact that this was a survey conducted on reddit results in certain amount of sample bias - the fact that 30% of the population comes from the software industry might not hold true if a survey was conducted more broadly instead.