[gwc]

That's just silly. A confidence interval by definition is a range; the commenter is arbitrarily choosing to post only the low end of that range. If the high end were posted instead, Washington's lead would look that much more impressive.

That's not to say the math is useless; a better statement (which could have been supported by the data) would have been "The difference between CA and WA is less than the measurement margin of error given WA's small sample size".