1. Users notoriously have no idea what they want. Surveys may not be the best method to determine the effectiveness of an icon because youre asking the user to predict their future behavior (which most are happy to do, confidently). The only way to test an icons effectiveness, is to test icons in a real world environment.

2. We hear this story every time Apple releases a product, when Google updates Gmail, when Facebook gave us Newsfeed. An instant gut reaction to change is not indicative of long term success or failure. People "hating" the icons is meaningless - they havent even been released yet. Change is good, its what drives our ambitions forward. Don't survey people and let the results scare you into stagnating.

3. http://observatory.designobserver.com/feature/graphic-design...

Its interesting the title says people hate the new icons, when they only preferred the old icons. Theres a comic to explain this situation quite well. http://theoatmeal.com/pl/state_web_winter/facebook_layout

Where does the "70%..." figure come from? It's not mentioned anywhere in the article. For that matter, the article itself throws around a lot of phrases like "clearly performing better", "proof", and "worst way", but the only evidence seems to be self-selected, explicit choice online polls?