Because we have no way of judging the bias of the poll takers or if they're even taking an adequate sample of an opinion. You can't even argue that 5,000 women took that poll. What if a man (or a large number of them) took the poll? What if a single developer troll took the poll with the sole intent of if swaying the data 5000 times? Did the analysis at That's Life! take into consideration possible corruption of their data, what steps did they take to prevent it? We don't know, we have to take the poll taker's evidence as unbiased and true. This is inherently unscientific. The output of the poll is not validated or cross referenced with any other credible studies in the article aside from hear-say from other online polls. The poll should be ignored by feminists, and non-feminsts alike due to the fact it is not rationally persuasive in the context provided by The Scotsman.