A New York Times article says that Donald Trump is "in a statistical tie" with Jeb Bush. Trump has 17%. Bush has 14%. The difference is within the margin of error for the Suffolk University / USA Today poll.
What's interesting to us is how the data has been cherry picked by the media. Consider these three different headlines, all based on the same survey:
- Donald Trump, Jeb Bush on the rise in new national poll on 2016 GOP field
Is Trump "on top," "tied" or "on the rise"? Perhaps it depends on your perspective. (It also depends on statistical significance - was the sample size large enough, how did they calculate the margin of error, etc.).
Oh, and another example of cherry picking - 30% of Republicans are undecided - roughly the same percentage who named Trump and Bush combined.
But that doesn't make the headlines.