American Pharoah for SI Sportsman of the Year

[GenuineRisk]

American Pharoah easily dominates his human competition and wins the readers' choice for Sportsman of the Year. The breakdown of voting by region, age and gender is interesting:

http://www.si.com/sportsmanofyear/20...sportsman-year

[Dunbar]

Originally Posted by GenuineRisk

American Pharoah easily dominates his human competition and wins the readers' choice for Sportsman of the Year. The breakdown of voting by region, age and gender is interesting:

http://www.si.com/sportsmanofyear/20...sportsman-year

Something's not right in those voting breakdowns. In the gender breakdown graph, KC appears to have twice as many male votes as Pharoah, while women voters appear to be split evenly between KC and Pharoah. So KC appears to have won. Yet in the overall tally, AP has 50% more votes. That makes no sense.

Okay, I see what's wrong. The gender and age breakdowns only included about 43,000 votes for all the nominees. But there were about 589,000 votes. So those gender and age graphs only show about 7% of the votes!

Somebody got carried away making pretty graphs but forgot about having it make sense or be representative of the data.

[bare it all]

Serena Williams ends up winning the actual title of sportsman of the year.

Take out the American Pharoah aspect and this selection still rubs me all the wrong way. Yes, Williams had another career year and did just what everyone expected her to do - except win the US Open in which she lost to an unknown Italian player in the semi-finals. Not even the finals. She was supposed to dominate and didn't and acted all put out like she usually does. She's a horrible sportsman and always has been. Women's tennis is notoriously weak at the top save Williams. She dominates an easily dominatable division. Sticking with tennis, a glance at the year Novak Djokovic had and anyone with a basic understanding of sports and competition can see that what he put together in 2015 was stunning. It's been called the "best year of tennis ever" in men's tennis. He was a set away from a Grand Slam (losing to Stan Wawrinka in the Aus Open final). He's currently miles ahead of the 2nd ranked player. I can't speak to the other nominees as I don't follow basketball or women's boxing or soccer or whatever else was nominated, but I can tell you the pick of William's in just the realm of tennis is an absolute sham because there were other competitors with stronger resumes for this year's award.

[GenuineRisk]

Originally Posted by Dunbar

Something's not right in those voting breakdowns. In the gender breakdown graph, KC appears to have twice as many male votes as Pharoah, while women voters appear to be split evenly between KC and Pharoah. So KC appears to have won. Yet in the overall tally, AP has 50% more votes. That makes no sense.

Okay, I see what's wrong. The gender and age breakdowns only included about 43,000 votes for all the nominees. But there were about 589,000 votes. So those gender and age graphs only show about 7% of the votes!

Somebody got carried away making pretty graphs but forgot about having it make sense or be representative of the data.

It's possible they pulled from the people who identified age and gender (I can't remember if I did, or had to), and not the rest who didn't. It's also possible they only analyzed a segment of the total data, but if they did, it's probably pretty representative of the total voting population.

Either way, I'll forward it on to my brother, who has a PhD in statistics and ask him what he thinks of the accuracy of the presented data. He lives for this kind of thing.

[Dunbar]

Originally Posted by GenuineRisk

It's possible they pulled from the people who identified age and gender (I can't remember if I did, or had to), and not the rest who didn't. It's also possible they only analyzed a segment of the total data, but if they did, it's probably pretty representative of the total voting population.

Either way, I'll forward it on to my brother, who has a PhD in statistics and ask him what he thinks of the accuracy of the presented data. He lives for this kind of thing.

I'm sure your brother will come to the same conclusion I did. The data shown in the the male/female graph totally misrepresents the overall data. Likewise, but less glaringly so, the data in the "by age" graph also misrepresents the overall data.

Look at it again--don't you agree that the male/female graph seems to show that KC had more votes than AP?

[GenuineRisk]

Originally Posted by Dunbar

I'm sure your brother will come to the same conclusion I did. The data shown in the the male/female graph totally misrepresents the overall data. Likewise, but less glaringly so, the data in the "by age" graph also misrepresents the overall data.

Look at it again--don't you agree that the male/female graph seems to show that KC had more votes than AP?

I don't; what I took from it was that AP's win was dependent on women voting for him in much larger numbers, percentage-wise than men did. That said, graphs can be notoriously poorly designed, you're right about that. There was one recently used on FOXNews that was cited as a textbook example of a graph, while not exactly contradicting the information, was drawn as to lead to an incorrect conclusion about it. It was pretty funny to look at.

The other thing is that stood out to me with the SI stuff is that it lists more women than men voting, which, not to generalize, but I can't help but find unlikely for a sports magazine. So I'm curious if they took a random sample of all voters, or if more women voters than men were willing to identify themselves by gender when voting (though, again, I can't remember if I was asked to categorize myself).

[GenuineRisk]

It will be interesting to see if the Zayat response (their tweets quoted by Bloodhorse are, er, a bit bitter) will garner them the same level of vitriol from racing fans as California Chrome's connections got after mouthing off after Chrome's Belmont loss.

http://www.bloodhorse.com/horse-raci...an-of-the-year

[Danzig]

the horse don't care...so who cares?

but...if they didn't want to give to a non human, why allow voting on a non human?

[Dunbar]

Originally Posted by GenuineRisk

I don't; what I took from it was that AP's win was dependent on women voting for him in much larger numbers, percentage-wise than men did.

It's true that a higher percentage of women than men voted for AP. But it's also crystal clear from those graphs that more people overall voted for Kansas City than for AP. If you only had those "by gender" graphs to go by, there's no rational way you would conclude that AP got the most votes.

Whatever way they got the sample of voters in that gender graph, it wasn't representative of the overall voting. There's nothing subtle about the discrepancy:

http://www.si.com/sportsmanofyear/20...sportsman-year (about half-way down the page, just above the comments)

[GenuineRisk]

Originally Posted by Dunbar

It's true that a higher percentage of women than men voted for AP. But it's also crystal clear from those graphs that more people overall voted for Kansas City than for AP. If you only had those "by gender" graphs to go by, there's no rational way you would conclude that AP got the most votes.

Whatever way they got the sample of voters in that gender graph, it wasn't representative of the overall voting. There's nothing subtle about the discrepancy:

http://www.si.com/sportsmanofyear/20...sportsman-year (about half-way down the page, just above the comments)

I wouldn't say "crystal clear," but I would agree that the bar graphs are a muddled mess- without showing the numbers next to the colors, there's not much you can take from the bar graphs (though they are very pretty). My brother's take on it was that it was highly unlikely it was representative data- if you're only taking respondents who self-identify their gender and age, you're not going to get a clear representation, because only using voters who voluntarily give out that data is already affecting the results. He mostly focused on the fact that the poll is clearly skewed by its nature to America, which made me laugh. To quote him:

"The world figure shows it was a pretty biased selection with the US dominating the number of respondents. The rest of the world cares about soccer and almost exclusively soccer. That figure does show that we do have to be careful in interpreting the data. Australia voted for Stephen Curry? Really? The American basketball player. Either there is a Rugby or Cricket player also named Stephen Curry or the only vote from Australia was by an American ex-pats living there."

But hey, props to Uruguay, for voting for the Pharoah!