Twice now you have claimed that this poll was not taken of likely voters, or even of registered voters. Can you tell us where your information came from, since it directly conflicts with what the article says?: "This survey was conducted between April 19-25, 2002 of 546 registered California voters likely to vote in November."
Field calls households at random and asks the person who answers if they are registered to vote and if they are likely to vote. Now most people -- from a good citizen standpoint -- will not admit that they aren't registered or that they hardly ever vote.
More sophisticated political pollsters call names from the voter registration database and they categorize how likely the person is to vote based upon how often they are recorded as voting in recent elections.
As I said, by leaving it up to the person who answers the phone to determine if they are registered and likely to vote, it becomes a poll of ALL ADULTS -- maybe not even that. I suppose a child with a deep voice could answer too.