Reply #64: What do you mean they have no margins of error [View All]

[Terry_M]

It's still a poll, and in this case you are depending on 2 things being accurate. That a) your sample is an accurate representation of party affiliation b) your sample is an accurate representation of the way people voted. Unless you poll everyone exiting, you will have a margin of error.
Also, you tried to turn the issue sideways. It's not whether independents or republicans support Obama. My actual question was is it safe to say that democrats have so far preferred Clinton, or are the results so close that there is no statistically significant difference. Again, see paragraph above for why the data MUST still have a margin of error.