To my knowledge, I don't think cell phone numbers are listed in phonebooks (I could be wrong). So the people being surveyed would only have landlines. That leaves out a big chunk of the population that have cell phones. That's one major bias.
Another bias is that the very long questions would make people hang up if they are in a hurry or find the topic boring. Those still willing to answer the questions are probably very passionate about the subject to skew the results one way or another. This happens in a lot of political surveys.
To fix these errors, the researchers need a way to contact anyone in that city and not just through the phone book. Perhaps they could have people put their names in a hat to be selected. Also, the questions should be shortened down to account for people's busy lives. The questions should be as neutral and objective as possible.