🏅 Top traders

Any poll? Even one which is excluding Trump? Perhaps that is not showing Haley leading Trump?
(This is listed on 538 NH polls page.)

@ChristopherRandles Needs to show Haley leading Trump. And the poll has to be one that is used for the 538 rolling average which this one is not.

A 538 poll shows Haley ahead. Resolved?

@MichaelLomuscio That's the one we've been debating. It's not clear if there's enough evidence that Haley is winning, or if it's only a tie.

@MichaelLomuscio 538 shows them tied, not with Haley ahead. Haley is listed first because of alphabetic sorting.

I’ve decided that the 40-40 ARG poll is not sufficient to resolve YES. It needs to be beyond doubt that Haley surpassed Trump in the number of respondents.

@thepurplebull The objectively right call 👍🏼

@Magnus any poll was the question.

@LandanSmock Market author should have made clear it needs to be a statistically significant margin then.

@LandanSmock Look, if you can get crosstabs data from ARG proving Haley got more respondents than Trump, I will resolve YES. Otherwise, I think it’s clear there has not been a single poll on 538 where Haley won.

@ClaytonNall Statistical significance is of no importance here. That is implied in the description which has not been changed.

@thepurplebull It definitely is of importance. If she is in the margin of error then you can’t say she is ahead

I’m a sucker for these types of markets but I miss the mark every time…

I think if ARG puts out another poll, it’ll show Haley leading, but otherwise this is NO

@MarkHamill ARG won’t. They’ve gotta know that they’re doing something wrong. Trump is going to win handily, and showing a poll with Haley winning will severely tarnish their credibility.

Pretty interesting table from the latest poll

This means that likely voter polls are overestimating the Haley vote. It also means that the specific methodoly the pollster uses to determine "likely voters" really matters.

@Shump That would reflect what was reported in Iowa, that enthusiasm is stronger for Trump primary voters

@Spice_N_Wolf I can't remember the source, but I read that half of Haley's voters said they would vote Democrat in the next elections, which is pretty insane

Crosstabs seem to imply Haley +0.5, but not for certain, and exact numbers aren't released.

@riverwalk3 Assuming they round to the nearest integer for crosstabs, the first crosstab upper bounds Trump at 40.09 and the male/female crosstab lower bounds Haley at 39.82. But that's all I can be certain about.

@riverwalk3 Obviously I’m biased, but since the poll doesn’t make those calculations (you are speculating), it shouldn’t resolve yes. The poll states 40-40, nothing more, nothing less.

@riverwalk3 thanks for pointing out the cross tabs, it definitely looks like Haley edged out Trump in this poll.

@FoxKHTML The description of the market does allow for checking the underlying data in case of a tie.

I don't think this should resolve YES unless we have the raw data. It's possible to get such cross-tabs even if Trump got more of the vote.

@Shump At the edge case (assuming rounding to the nearest integer), Trump got at most 239 out of 600 and Haley got at least 239 out of 600. Haley most likely got more, but this is all I could prove.

@riverwalk3 Thanks for checking, I guess we can't resolve yet then.

@TimothyJohnson5c16 This should only be invoked if 538 lists it as 40-40 but the poll specifics explicitly state (eg) 40.3-39.8.

@FoxKHTML like this?

@Spice_N_Wolf That shows 40-40…?

@FoxKHTML Right, which means the following is invoked:

"A poll showing both candidates at the same percentage may be evaluated further at the original source."

Which means cross tabs can be used, which is also why I believe 538 put Haley above Trump in the listing.

@Spice_N_Wolf I think they put her ahead due to alphabetical order

@MarkHamill I checked for that based on older polls where there was a tie and found it wasn't alphabetical

@Spice_N_Wolf The pollster didn’t state that Haley was ahead, though. It’s speculative. That’s my point. So, it cannot resolve.