For a better browsing experience, please upgrade your browser.

You’re reading Significant Digits, a daily digest of the telling numbers tucked inside the news.

1 paid staffer

Former Texas Gov. Rick Perry’s presidential campaign appears to be in dire straits: His polling numbers are negligible (they’re in rounding-error territory) and according to reports from Perry aides, he’s down to a single paid staffer in the entire state of Iowa. Basically barring a huge turnaround, Perry is that kid from “The Hunger Games” who fell for the ol’ “big pile of shiny weapons” trick in the first five minutes. [USA Today]


2.1 percent

Increase in revenue collected in Colorado from alcohol excise taxes from June 2014 through May 2015. Reminder that Colorado legalized marijuana, so this is good news for a local alcohol industry worried that cannabis would cut into its market. So far at least, marijuana has not stopped the rest of the intoxicant industry from expanding. I don’t normally do calls to action here, but you really need to read this story if only to see the greatest use of the idiom “a high tide lifts all boats.” [The Guardian]


11 turnovers

Ladies and gentlemen who support the Washington NFL team, please allow me to re-introduce your starting quarterback: Kirk Cousins, who had 11 turnovers in five games last year. Apparently head coach Jay Gruden elected to sit anticipated starter Robert Griffin III for the first game of the year in favor of Cousins, despite urging from management to play RG3. [ESPN]


23 percent

According to a new poll, the support among likely Iowa caucus-goers for GOP candidate Ben Carson, a neurosurgeon. Carson is tied with Donald Trump. [Monmouth University]


29 districts

For the first time ever, Singapore will have a two-or-more party contest in each of its 29 districts in an upcoming election. The politics of the city-state has been dominated by the unbroken rule of the People’s Action Party, but following the demise of the party’s longtime leader in March, other parties have seen an opportunity to compete. [Bloomberg]


75 percent

A proposed takeover bid of third-party travel booking site Orbitz by competitor Expedia could result in a company that controls about 75 percent of the domestic market in the field, according to research from Phocuswright. This potential deal has put hoteliers on notice, with many trying to find legal ways to attract and keep customers with factors other than price. [The New York Times]


80 percent

Efficiency of a Swiss “solar sunflower” design for solar energy collection. The apparatus, which kind of looks like a satellite dish hooked up to a disco ball, is able to produce 12kW of electricity and 21kW of thermal energy. It’s still pricier per watt than cheap solar cells, but its efficiency is intriguing — and it’s pretty to look at. [Ars Technica]


$100 gift cards

Jeep really, really wants to get recalled vehicles into the shop, and to make sure owners get around to bringing in their cars for necessary repairs, the company has resorted to giving out $100 gift cards. [Bloomberg]


500 calorie limit

A New York City council member wants to cap the nutritional intake on fast food children’s meals at 500 calories. I can see an argument in favor of that, sure, but Council Member Benjamin J. Kallos also wants to mess with the macronutrient composition of food by capping fat content. The last time the government tried to pull that off, things went pretty badly. [New York Observer]


1,279 career rushing attempts

The NFL’s oldest running back, Fred Jackson, was cut from the Buffalo Bills Monday after almost a decade with the team. Jackson, a fan favorite, was the top rusher for Buffalo last year. “F—ing Bills,” added local political reporter and Buffalo fan Harry Enten. “If they go 7-9 and cut Fred Jackson I will be so pissed,” Enten added in a late-night message. [ESPN]

If you haven’t already, you really need to sign up for the Significant Digits newsletter — be the first to learn about the numbers behind the news.

Comments Add Comment

A few weeks ago, I wrote a piece headlined “The Bernie Sanders Surge Appears To Be Over.” The article (and, more emphatically, the headline) argued that Sanders’s rapid ascent in the presidential primary polls had slowed or stopped in polls of New Hampshire and Iowa, at least based on the polls we were seeing at that time. Ever since then, Bernie Sanders fans have been tweeting at me — sometimes kindly, sometimes not — after every new poll that showed their candidate doing well.

And guess what? They sorta have a point. Although his gains may not be as great as before, polls in August showed Sanders continuing to pick up support in New Hampshire (the situation in Iowa is less clear). But the fundamentals of the race haven’t changed very much. In particular, Sanders has shown little sign of winning over votes from African-Americans or Hispanics, which would limit his growth as the race moves on to more racially diverse states.

First, let’s talk a bit more about his recent gains in the polls, starting in New Hampshire. In the original article, I wrote that “Sanders rose from June to July in the Granite State, but his ascent slowed.” But Sanders picked up the pace in August, and support for Hillary Clinton has slowly dropped, according to live-interview1 polls.2 Here are the monthly averages:

IOWA NEW HAMPSHIRE
MONTH CLINTON SANDERS CLINTON SANDERS
August 47% 27% 37% 44%
July 49 25 42 34
June 51 29 47 30
May 57 16 62 18

While the Sanders surge continued in New Hampshire, the polling in Iowa is more ambiguous. There was a lot of coverage of this weekend’s Des Moines Register poll (produced by top-notch pollster Ann Selzer), which found Clinton leading Sanders by just 7 percentage points, 37 percent to 30 percent. But Sanders’s numbers in Iowa, even in the Des Moines Register survey,3 have remained relatively consistent.

Clinton, on the other hand, continues to bleed support. The Des Moines Register poll was a bit of an outlier in terms of where it showed Clinton, but Selzer is so good at polling Iowa, you would be a fool to dismiss it.

So, Clinton’s margin over Sanders is at its lowest in both Iowa and New Hampshire. And while national polls weren’t the focus of my initial article, Sanders has also closed in on Clinton in those.

So why do I still think Sanders is a factional candidate? He hasn’t made any inroads with non-white voters — in particular black voters, a crucial wing of the Democratic coalition and whose support was a big part of President Obama’s toppling of Clinton in the 2008 primary. Not only are African-Americans the majority of Democratic voters in the South Carolina primary (a crucial early contest), they make up somewhere between 19 percent and 24 percent of Democrats nationwide. In the past two YouGov polls, Sanders has averaged just 5 percent with black voters. Ipsos’s weekly tracking poll has him at an average of only 7 percent over the past two weeks. Fox News (the only live-interview pollster to publish results among non-white voters in July and August) had Clinton leading Sanders 62-10 among non-white Democrats in mid-July and 65-14 in mid-August. Clinton’s edge with non-whites held even as Sanders cut her overall lead from 40 percentage points to 19.

There are other indications that Sanders is unlikely to win the nomination. He hasn’t won a single endorsement from a governor, senator or member of the U.S. House of Representatives (unlike Obama at this point in the 2008 campaign). Sanders is also well behind in the money race (again, unlike Obama). These indicators haven’t changed over the past month.

But even if you put aside those metrics, Sanders is running into the problem that other insurgent Democrats have in past election cycles. You can win Iowa relying mostly on white liberals. You can win New Hampshire. But as Gary Hart and Bill Bradley learned, you can’t win a Democratic nomination without substantial support from African-Americans.

Footnotes

  1. Inclusion of non-live-interview polls does not change this analysis. ^
  2. If a poll tested multiple primary matchups — with Joe Biden and without Biden — I used the question with Biden. ^
  3. His 30 percent in the Des Moines Register poll was just south of the 33 percent he earned in a June Quinnipiac University poll and the 31 percent he took in a CNN poll from earlier this month. ^

Comments Add Comment

One question we’ve asked about Donald Trump’s campaign is what would happen if another candidate were to get a surge of media coverage, throwing a wrench in the works of Trump’s perpetual attention machine. Based on the latest polls, we may soon get the best test of this question since Trump entered the race.

The potential surger is Ben Carson, the much-awarded neurosurgeon who is running for political office for the first time. Carson is near the top of the Republican field in two new polls of Iowa. A Monmouth University poll has Carson with 23 percent of the vote and tied for the lead with Trump, while the Des Moines Register’s poll has Carson in second place, with 18 percent of the vote to Trump’s 23 percent. Carson has also been gaining ground in national polls and is in second place behind Trump, according to Huffington Post Pollster’s averaging method.

Carson’s gains in the polls have come organically — in fact, he’s been receiving remarkably little media attention. From June 28 through Aug. 20, Carson received only 0.9 percent of the news coverage of the top 164 Republican candidates, based on the number of Google News “hits” he received. That puts him at 14th in the field. Trump has received about 60 times more media coverage than Carson.

silver-datalab-carson-1

While it can be foolish to predict what happens to the polls in the short run, there’s a pretty obvious case to be made that Carson is on an upswing as part of a “discovery, scrutiny and decline” polling cycle of the sort that Newt Gingrich and Herman Cain (among others) experienced in 2011. If Carson’s doing this well with so little media attention, imagine what happens when he gets some. Polls will trigger more coverage of Carson’s campaign, which will in turn improve his standing in the polls, which will produce yet more coverage, and so forth.

Carson also has outstanding favorability ratings among Republicans, which could give him more room to grow. And it’s not as though he’s a dull story to cover. While Carson is more mild-mannered than Trump — and possibly a lot smarter — he, like Trump, has a history of stoking controversy through impolitic statements.

The question is what happens to Trump’s numbers when Carson surges. (Or if Carson doesn’t, when another candidate like Ted Cruz inevitably does some weeks or months from now.) If Trump is more like the Gingriches and Cains of the world, his support may erode pretty quickly once there’s another GOP “flavor of the month” who appeals to voters seeking an outsider to mainstream politics. An alternative possibility, however, is that Trump is more like a Pat Buchanan or Ron Paul — a factional candidate who is relatively immune from shifts of opinion elsewhere in the Republican field, but also has a low ceiling on his support. Either way, Trump is not very likely to win the Republican nomination — and neither is Carson — but we’ll learn something about the nature of his support.

Footnotes

  1. Inclusion of non-live-interview polls does not change this analysis. ^
  2. If a poll tested multiple primary matchups — with Joe Biden and without Biden — I used the question with Biden. ^
  3. His 30 percent in the Des Moines Register poll was just south of the 33 percent he earned in a June Quinnipiac University poll and the 31 percent he took in a CNN poll from earlier this month. ^
  4. Sorry, Jim Gilmore — we didn’t collect data on you. ^

Comments Add Comment

You’re reading Significant Digits, a daily digest of the telling numbers tucked inside the news.

1 year

A team of six NASA astronauts began a one-year isolation project in Hawaii to simulate a year on Mars without fresh air or privacy. I really hope Fox options this for a terrible reality show. [BBC]


3 times as likely

A new study found that British teenagers who identified with the Goth subculture at age 15 were three times as likely to be clinically depressed in their teen years as those who did not. [The Guardian]


4 months

An Australian man was released from a months-long detention after a substance in his possession originally believed to be methamphetamine turned out to just be Epsom salts. He was in the clink for four months before police finally got around to making sure the rocks were meth. If I were him, I would be so mad at the guy who sold me that “meth.” [The Independent]


72.2 percent

Percentage of pop stars — a set of the 50 top-selling artists of each decade — from 1950 to 2014 who were born in the United States. The U.K. comes in at a far second, with 15.6 percent of the stars in the set. [CityLab]


118 years

President Obama approved an order from the Secretary of the Interior to rename the peak known as Mount McKinley for the last 118 years to its original Alaskan appellation, Denali. This has been an Alaska vs. Ohio fight for some time: the former — where the mountain is in fact located — has long fought to call the peak by its Koyukon Athabascan name, while the latter, where President William McKinley was born, has sought to nix the change. Obama — a guy who presumably gets non-contiguous states, being from Hawaii — finally approved the change. Also, Speaker John Boehner is ticked. [Alaska Dispatch News]


11,000 years

A wooden statue found in the peat bogs of Russia 125 years ago has just been dated to 11,000 years old, making it the oldest wooden artifact in the world. It’s covered in an encrypted code nobody can decipher, but scholars seem to think it contains the outlines of a belief system. [Yahoo News]


$108,000

About 850 vintage Atari games found in a landfill in New Mexico have sold for a total of $108,000 on eBay. The city of Alamogordo will get most of the funds, with $16,000 going to the local historical society. [CTV News]


$500,000

How much Major League Baseball paid for the domain MLB.com in 2000, in addition to six months of free advertising for the domain’s original owners, Morgan, Lewis & Bockius LLP. It’s been a steady fight in the following 15 years to nail down domain names for each of the league’s 30 franchises, and only three preferred domains remain out of MLB’s grasp: Giants.com, which is owned by the New York football team, Rays.com, owned by a Seattle restaurant, and Twins.com, which is a whole other story involving a pair of legitimate twins who refuse to sell. [Grantland]


$1.8 million

“We Are Your Friends,” a movie about electronic dance music starring Zac Efron, had one of the worst wide-release openings ever, earning only $1.8 million from 2,333 screens this past weekend. [Variety]

5.5 billion barrels

A huge natural gas field has been found off the coast of Egypt by an Italian energy group. It’s got up to 30 trillion cubic feet of gas, about the same as 5.5 billion barrels of oil. [BBC]


If you haven’t already, you really need to sign up for the Significant Digits newsletter — be the first to learn about the numbers behind the news.

Footnotes

  1. Inclusion of non-live-interview polls does not change this analysis. ^
  2. If a poll tested multiple primary matchups — with Joe Biden and without Biden — I used the question with Biden. ^
  3. His 30 percent in the Des Moines Register poll was just south of the 33 percent he earned in a June Quinnipiac University poll and the 31 percent he took in a CNN poll from earlier this month. ^
  4. Sorry, Jim Gilmore — we didn’t collect data on you. ^

Comments Add Comment

This is The Week In Data, our data journalism roundup. Here you’ll find the most-read FiveThirtyEight articles of the past week, as well as gems we spotted elsewhere on the Internet.

MOST READ

  1. I’m Better Than 245 NFL Players At Catching Footballs
  2. Worried About The Stock Market? Whatever You Do, Don’t Sell.
  3. The Happiest And Saddest Fans In Baseball
  4. Donald Trump Is Running A Perpetual Attention Machine
  5. The Best Jobs Now Require You To Be A People Person
  6. Most Of The Biden Speculation Is Malarkey
  7. Katrina Washed Away New Orleans’s Black Middle Class
  8. When Trump Attacks!
  9. Public Transit Should Be Uber’s New Best Friend
  10. Don’t Read Too Much Into Ohio State’s Unanimous No. 1 Ranking

ELSEWHERE ON THE INTERNET

Economics is expensive: It’s not news to students that college textbooks are expensive, but how do different majors compare on the cost of materials? This article covers everything from price changes for textbooks over time to the average price of new vs. used books. Essential reading. [Priceonomics]

Screen Shot 2015-08-30 at 9.22.36 AM

Dystopian dating: Just 5.5 million of the roughly 37 million accounts in recently leaked database of the online dating site Ashley Madison belonged to women. Take a closer look, and it appears that many of those female accounts on the website for having a affair weren’t active, or used at all. After looking at the evidence, Annalee Newitz reckons, “The overwhelming majority of men using Ashley Madison weren’t having affairs. They were paying for a fantasy.” [Gizmodo]

1403242054690450350

Counting the pennies: Artist blake Fall-Conroy wants to create socially conscious work, and his “Minimum Wage Machine” (2008-2010) is also an elegant work of data visualization. (While it’s a few years old, I spotted it this week after Nathan Yau mentioned it — and it was too good to omit.) Turn the crank to get a penny every 4.11 seconds. Keep it up and you’ll earn $8.75 an hour, the New York state minimum wage in 2015. [Blake Fall-Conroy]

Minimum-wage-machine

Asylum applicants: As the crisis in Calais deepens, the debate about migration in Europe has turned to those who are seeking asylum. This article summarizes asylum applications and decisions made in European Union countries last year. [The Guardian]

Screen Shot 2015-08-28 at 10.01.26 AM

More than one mass shooting per day in 2015: This week’s fatal shooting in Virginia was (as of Friday) the 247th U.S. shooting in 2015 to involve four or more victims, according to an online group that tracks mass shootings. [Wonkblog]

imrs.php

Footnotes

  1. Inclusion of non-live-interview polls does not change this analysis. ^
  2. If a poll tested multiple primary matchups — with Joe Biden and without Biden — I used the question with Biden. ^
  3. His 30 percent in the Des Moines Register poll was just south of the 33 percent he earned in a June Quinnipiac University poll and the 31 percent he took in a CNN poll from earlier this month. ^
  4. Sorry, Jim Gilmore — we didn’t collect data on you. ^

Comments Add Comment

Another day, another batch of mostly redundant and anonymously sourced stories about whether Vice President Joe Biden will run for president. Some of those stories, however, are getting ridiculous. So FiveThirtyEight’s politics writers met in Slack to pick over the latest Biden coverage, our own assumptions and the state of the 2016 Democratic primary. This is an edited transcript of the conversation.


micah (Micah Cohen, senior editor): So, the will he/won’t he speculation about Joe Biden hasn’t slowed down, but do either of you buy the argument that a Biden run could actually help Hillary Clinton?

hjenten-heynawl (Harry Enten, senior political writer): I don’t think it would be particularly helpful to Clinton. Forget about all the BS about whether Clinton runs better when she’s in trouble. Personally, I never got that. If she were so good at running when she was in trouble, then why did she lose in 2008?

Rather, why would Biden run? Sure, he’s in his 70s and this is his last shot, but he also has a family to take care of. He’d likely only run if he concludes he has a better than nominal chance of winning. And that conclusion would be quite different from what the current metrics, such as endorsements, suggest. Biden may have an insight on the invisible primary that isn’t visible to the rest of us.

natesilver (Nate Silver, editor in chief): The irony is that the media has exaggerated all sorts of threats to Clinton, who remains in good shape for the nomination. But then you have the one thing that would be a tangibly bad sign for her campaign — the vice president of the United States running for the nomination against her! — and there are lots of “smart takes” about how it could help Clinton.

hjenten-heynawl: What we’ve argued this entire time is that Sen. Bernie Sanders has a weakness among the party actors (i.e., he doesn’t have any endorsements), and that he has no longtime connections to the Democratic Party (remember, he’s not a Democrat). Biden, on the other hand, has been in major federal office in Washington since 1973. He’s someone who could conceivably reach out to all members of the party. He’s already polling better among African-Americans than Sanders, for instance.

micah: Let’s break this down a little: Both of you seem to think Biden entering the race is inherently bad for Clinton — he’d be the most serious competition for the nomination she’s faced. But would there be a couple side benefits, like that by giving the media a horse race to cover, there would be less focus on Clinton’s scandals?

natesilver: Well, first of all, it’s not just that Biden would be a more formidable competitor to Clinton than Sanders. I don’t know that Biden would be all that great a candidate, in fact. But Biden running would signal that concern about Clinton among Democratic Party elites had gone from the bedwetting stage to something more serious.

micah: Is bedwetting not serious?

hjenten-heynawl: I mean, it depends how old you are.

natesilver: But the other big problem (as we and others have pointed out before) is that Biden doesn’t have much rationale to run other than if Clinton has “trust”/scandal problems. He might never come out and say it, but that would be the whole basis for his campaign. They don’t really differ in any meaningful way on policy.

micah: But your logic seems circular: “Biden will only enter the race if Clinton is in big trouble, and therefore if Biden enters the race it means Clinton is in trouble.” What if all the party actors are telling Biden that he shouldn’t run, that they’re backing Clinton, and Biden just wants to run? It’s his last chance. And he enters the race.

natesilver: What I’m saying is that there’s a lot of information we’re not privy to, about what Democratic elites are thinking. Sure, there’s some reporting on it, but a lot of that reporting needs to be looked at skeptically — like because it relies on anonymous sourcing, or cherry-picked information from a media that would like to make the race seem competitive. The one tangible sign we have about what Democratic elites are thinking — endorsements — looks really good for Clinton. But Biden running would be a tangible sign too.

Contra Maureen Dowd or whatever, this isn’t necessarily a personal decision for Biden, or at least not entirely one. He’s a party guy. He’s the vice president. He’s not likely to run unless he thinks it’s in Democrats’ best interest.

hjenten-heynawl: Endorsements are merely a proxy for intra-party support. And proxies are wrong from time to time. They’re imperfect. And I don’t buy Biden is desperate to run. He reportedly indicated this week in a phone call with Democratic National Committee members that he and his family are grieving. The man lost his best friend and son. He wants to be there for his family. If I lost my father (my best friend), I don’t take off running for president just because I feel like it. I run because I think I can help my party, and because I think I can win.

natesilver: Right. It’s possible that Biden assesses the problem and miscalculates. But running for president would be a calculated decision on his behalf.

And, by the way, if you read the reporting on Biden carefully, it suggests that the decision is very, very calculated. He’s taking as long as possible to decide whether to enter — and at a time when it’s already pretty darn late to begin a campaign — because he wants to collect more information on whether Clinton’s in trouble or not.

hjenten-heynawl: BINGO. He’s meeting with a ton of people who represent different wings of the party, such as Sen. Elizabeth Warren and Richard Trumka (head of the AFL-CIO). He’s doing that, one would think, because he wants to understand what they are seeing. What are their people, their constituents, telling them.

micah: OK, so let’s say Biden gets in. The night before he announces, he’s sitting with his family and some advisers and they’re talking about why they can beat Clinton (based on everything they hear during these weeks of meetings). What are they saying? Does it all come down to email/scandal? Or would they be pointing to something else in the Clinton campaign or electorate? (I want data.)

natesilver: If you want data, and Biden’s camp is looking at the same data, then they shouldn’t be running in the first place. Unless they think the scandal will be Clinton’s undoing.

Clinton remains extremely popular with Democrats, and that popularity is pretty broad-based. White liberals might not like her as much as white moderates, Hispanics, or African-Americans, but as we’ve argued before, their support for Sanders is more an indication that they like him than that they dislike Clinton.

Some of the reporting around what Biden’s coalition would be doesn’t make any sense. See, for example, from Politico:

Biden’s circle has identified what they see as their potential voting blocs: Reagan Democrats, Jews, an LGBT base that largely credits him with pushing President Barack Obama into supporting gay marriage, and Rust Belt voters. They believe he’ll benefit from better stump skills than any of the other candidates running.

There’s no evidence that any of these groups are weaknesses for Clinton. Nor are they all that large, nor do they have very much in common.

micah: What about the Quinnipiac poll out this week showing Biden running better than Clinton against Republicans in general election matchups? And that voters don’t think Clinton is trustworthy or honest?

natesilver: I don’t think you can compare a declared candidate in Clinton — who’s been getting a ton of scrutiny from the press, some deserved and some not — against a hypothetical candidate who has a halo around him because the press would love to see a huge fight for the nomination.

Over the long run, Clinton’s favorability numbers have been no worse than Biden’s. Often a little better.

hjenten-heynawl: General election polls of candidates who aren’t running in the primary are ridiculous. Once he enters, all of Biden’s faults will be put on the table. And there are a lot to play with. If there weren’t, he’d have done better when he ran in past elections.

micah: From the WSJ writeup of the Quinnipiac poll:

The Quinnipiac poll found that 51% of voters have an unfavorable impression of her, her worst score ever on that measure. The poll also found that 61% of voters say she is not honest and trustworthy, another record low. On the honest and trustworthy question, that is up from 57% in a July Quinnipiac poll.

natesilver: Here’s the problem, Micah. Lots of people, political reporters especially, believe in momentum. If something goes from 50 to 45 percent, they assume it will keep going down, until it hits 40, 35, etc.

But empirically, the opposite is closer to being true. At least when it comes to polling.

If something goes from 50 to 45, it’s more likely to bounce back to 50 than to continue declining. Mean-reversion tends to be stronger than momentum. At least over the long term — the short term is sometimes a different story. But it’s the long term we should be concerned with, given that it’s still only August.

The Clinton who has a 42 percent favorability rating today isn’t really all that different than the one who had, I dunno, a 52 percent favorability rating at the start of the campaign, or a 48 percent favorability rating when she was running in 2008, or whatever. She is different than Clinton as secretary of state or first lady, because those are closer to being nonpartisan positions. So she can’t expect to see those numbers again, at least not while she’s a presidential candidate. But the odds are that her favorability ratings would revert to the mean by Election Day next year, which in her case means about 50/50.

hjenten-heynawl: Remember when there was talk about whether Chris Christie would get into the 2012 race? Or whether Fred Thompson would get into the 2008 race? Or Wesley Clark into the 2004 race? Those guys were tied or leading in the primary polling at the time. Biden’s best percentage so far has been 18 percent. He’s down nearly 30 percentage points to Clinton. Clinton is still in a ridiculously strong position.

natesilver: Yeah, I saw some article that offhandedly asserted Biden was polling exceptionally well given that he wasn’t in the race yet. Polling at 12 percent or 15 percent or 18 percent among members of his own party doesn’t seem that great to me for a guy who is vice president of the United States.

hjenten-heynawl: But we don’t have all the information. We believe Clinton is strong based on polling, money and support from party actors. If Biden were to enter, though, it says to us that he has a piece of information that we aren’t privy to. And this information is that Clinton is weak — for whatever reason. If he doesn’t enter, it’s a confirmation that she is strong within the party.

natesilver: Part of this is looking for verifiable evidence in an environment where the media has an interest in overrating how competitive the Democratic race is.

By most objective measures, Clinton is doing really well in the nomination hunt. About as well as any non-incumbent candidate has been doing up to this point in time. So, on the one hand, we look at that data and it makes us skeptical that Biden will convince himself to run. On the other hand, it means we have more reassessment to do if Biden in fact does run.

micah: OK, let’s say Biden gets in. How does he win? Does he come in guns blazing on email and trustworthiness? Does he claim the Obama mantle?

natesilver: How does he run or how does he win? I’d guess that his messaging would be rather cryptic at first. Because the way he wins is basically if Democrats decide that Clinton is too much of a liability because of her scandals. But Biden doesn’t want to come right out and say that. Debating Clinton on policy is also awkward, though, given that they have few real differences. And that, to the extent they do, one of them is going to be criticizing the Obama administration’s policy, which is an odd look for an incumbent party trying to win another term in office.

hjenten-heynawl: Let’s start with this: Clinton must perform disappointingly in the Iowa caucuses. If Clinton wins in Iowa by a convincing margin, this thing is going to take off. I don’t know how Clinton loses in Iowa, necessarily, but that’s where it needs to begin. Biden cannot wait until later states to take her on. Her money and momentum will be too great. So it’s Iowa or bust. Now, it could be that Sanders comes close in Iowa — it doesn’t have to be Biden, but he’s gotta do reasonably well.

natesilver: Yeah, I agree. I mean, one way Biden wins is if there’s some new scandal (or some new wrinkle to the email scandal) that’s so bad Clinton drops out. That’s sort of obvious, I suppose.

Short of that, it might come down to the timing. Say there’s some bad news for Clinton that drops a couple of weeks before Iowa. Iowa is taken as a referendum on her campaign, and she fails that referendum.

micah: And what happens to #feelthebern if Biden jumps in?

hjenten-heynawl: I think he continues on the path it was on. He’ll continue to get white liberals and that’s about it. I guess you could argue one way or another whether this slightly boosts his odds, but I think it doesn’t help him. If anything it could steal attention away from him as the anti-Clinton.

natesilver: Yeah, I don’t think Sanders’s support will be affected that much. At the margin, it might make it easier for him to win the plurality in a caucus state here or there. But Bernie will keep on Bernin’.

What I don’t think we’re likely to see is a case where the Clinton-Biden fight drags out for months and months, and then we’re all doing a bunch of delegate math, involving Clinton and Biden and Sanders, in May. As Harry said earlier, a Biden candidacy would either gain traction or collapse pretty quickly based on how it did in Iowa and New Hampshire.

hjenten-heynawl: Support for an anti-Clinton is either there or not.

micah: So on our initial question — “Could a Biden run help Hillary Clinton?” — we think the answer is: “No. And also, it probably wouldn’t affect Sanders much either.”

Is that right?

natesilver: A Biden run would be the worst news Clinton has had so far in the campaign. She’d still probably be the favorite, however.

Footnotes

  1. Inclusion of non-live-interview polls does not change this analysis. ^
  2. If a poll tested multiple primary matchups — with Joe Biden and without Biden — I used the question with Biden. ^
  3. His 30 percent in the Des Moines Register poll was just south of the 33 percent he earned in a June Quinnipiac University poll and the 31 percent he took in a CNN poll from earlier this month. ^
  4. Sorry, Jim Gilmore — we didn’t collect data on you. ^

Comments Add Comment

You’re reading Significant Digits, a daily digest of the telling numbers tucked inside the news.

1 voter

Gerrymandering can backfire, moneyed elites: A redistricting plan accidentally left a single University of Missouri student as the sole registered voter in a community improvement district in Columbia, Missouri. By herself, Jen Henderson, age 23, gets to vote on a sales tax increase for her area. [Columbia Daily Tribune]

2.45 percent

What a week! The Nasdaq Composite gained 2.45 percent Thursday, rebounding from a tumultuous couple of days when the market appeared to be in a correction. Aren’t you glad you listened to Ben Casselman? [CNBC]


20 percent

Percentage of Americans who talk about their religion online in a typical week. [Pew Research Center]


41 percent

According to a survey of 20,000 U.S. service-members, 41 percent of women in the military have suffered sexual harassment. About 4 percent of men surveyed also reported they had been sexually harassed. [Reuters]


71 percent

A Billboard survey of music executives found that 71 percent of respondents thought the music streaming service Tidal would survive for one more year or less. This isn’t very encouraging for the Jay-Z helmed app. Then again, if there’s a group on the planet to trust less about technology than music industry executives, I’d love to meet them and introduce them to fire, rounded stones and pointed sticks. [Billboard]


103 points

Consumer Reports is typically a no-nonsense evaluator of products. But that fact made the magazine’s initial score for the Tesla Model S P85D, 103 out of a possible 100, raise some eyebrows. The test is broken! Consumer Reports eventually had to tweak their model, but does this mean the blender I bought that I thought scored a 65 actually did worse? [Forbes]


152 socks

Current record for number of socks worn on one foot. That seems like a dubious mark to enshrine. Nonetheless, Guinness wants to record it, and more importantly, someone was possessed to attempt such a record. What drives someone to break a dumb record? The motivations are all across the board. [The Atlantic]


63 million people

Following protests, mobile Internet in the Gujarat state in India — home to 63 million — has been disabled and blocked. [The Next Web]


1 billion visitors

Facebook hit a milestone on Monday, founder and CEO Mark Zuckerberg announced, with 1 billion people visiting the service in a single day. While Facebook has about 1.5 billion active monthly users, this is the first time the site has had 1 billion unique people visit in a day. [Slate]


If you haven’t already, you really need to sign up for the Significant Digits newsletter — be the first to learn about the numbers behind the news.

Footnotes

  1. Inclusion of non-live-interview polls does not change this analysis. ^
  2. If a poll tested multiple primary matchups — with Joe Biden and without Biden — I used the question with Biden. ^
  3. His 30 percent in the Des Moines Register poll was just south of the 33 percent he earned in a June Quinnipiac University poll and the 31 percent he took in a CNN poll from earlier this month. ^
  4. Sorry, Jim Gilmore — we didn’t collect data on you. ^

Comments Add Comment

“There’s a difference between an answer and a result. But all the incentives are pointing toward telling you that as soon as you get a result, you stop.”

Subscribe: iTunes | Stitcher | SoundCloud | ESPN PodCenter | RSS


If you Google “is science broken?” you gather a surprising number of hits. It’s a question born of a perceived moment of crisis: Retractions are on the rise, the gap between what the public believes and what scientists believe is vast, and (perhaps relatedly) the number of scientists who say “this is a good time for science” continues to drop. The media doesn’t help. Headlines tend to overstate findings in search of catchy takeaways — drink eight glasses of water! — when the research is often much more nuanced.

On this week’s episode of our podcast What’s The Point, FiveThirtyEight’s Christie Aschwanden surveys the state of science and what it says about the impact research can have on our daily lives. She recently wrote an article addressing the question of whether science is broken, which also features an interactive graphic by Ritchie King on a particularly troubling technique of data manipulation known as “p-hacking.” Aschwanden concludes that science isn’t necessarily broken, but that it’s very, very hard — the way it always has been.

Plus, this week’s Significant Digit: In 2014, 218 newborns were named Anakin.

Stream or download the full episode above. Below is an excerpt from Christie’s story. And here’s the interview with Neil deGrasse Tyson referenced in the episode.


Science’s moment of crisis

Excerpt from Christie Aschwanden’s article for FiveThirtyEight. Read the full story.

After the deluge of retractions, the stories of fraudsters, the false positives, and the high-profile failures to replicate landmark studies, some people have begun to ask: “Is science broken?”

I’ve spent many months asking dozens of scientists this question, and the answer I’ve found is a resounding no. Science isn’t broken, nor is it untrustworthy. It’s just more difficult than most of us realize. We can apply more scrutiny to study designs and require more careful statistics and analytic methods, but that’s only a partial solution. To make science more reliable, we need to adjust our expectations of it.

Science is not a magic wand that turns everything it touches to truth. Instead, “science operates as a procedure of uncertainty reduction,” said [Brian] Nosek, of the Center for Open Science. “The goal is to get less wrong over time.” This concept is fundamental — whatever we know now is only our best approximation of the truth. We can never presume to have everything right.

“By default, we’re biased to try and find extreme results,” [John] Ioannidis, [a] Stanford meta-science researcher, told me. People want to prove something, and a negative result doesn’t satisfy that craving. Ioannidis’s seminal study is just one that has identified ways that scientists consciously or unconsciously tip the scales in favor of the result they’re seeking, but the methodological flaws that he and other researchers have identified explain only how researchers arrive at false results. To get to the bottom of the problem, we have to understand why we’re so prone to holding on to wrong ideas. And that requires examining something more fundamental: the biased ways that the human mind forms beliefs.


If you’re a fan of What’s The Point, subscribe on iTunes, and please leave a rating/review -- that helps spread the word to other listeners. And be sure to check out our sports show Hot Takedown as well. Have something to say about this episode, or have an idea for a future show? Get in touch by email, on Twitter, or in the comments.

What’s The Point’s music was composed by Hrishikesh Hirway, host of the “Song Exploder” podcast.

Footnotes

  1. Inclusion of non-live-interview polls does not change this analysis. ^
  2. If a poll tested multiple primary matchups — with Joe Biden and without Biden — I used the question with Biden. ^
  3. His 30 percent in the Des Moines Register poll was just south of the 33 percent he earned in a June Quinnipiac University poll and the 31 percent he took in a CNN poll from earlier this month. ^
  4. Sorry, Jim Gilmore — we didn’t collect data on you. ^

Comments Add Comment

Last week, I made the case that science isn’t broken, it’s just hard. Today in the journal Science, a group of more than 350 researchers published a paper demonstrating the challenges of a core step in scientific inquiry: replication.

The project, spearheaded by the Open Science Collaboration, aimed to replicate 100 studies published in three high-profile psychology journals during 2008. The idea arose amid a growing concern that psychology has a false-positive problem: In recent years, important findings in the field have been called into question when follow-up studies failed to replicate them, hinting that the original studies may have mistaken spurious effects for real ones.

“The idea was to see whether there was a reproducibility problem, and if so, to stimulate efforts to address it,” project leader Brian Nosek told me. In total, 270 co-authors and 86 volunteers contributed to the effort.

This wasn’t a game of “gotcha.” “Failing to replicate does not mean that the original study was wrong or even flawed,” Nosek told me, and the objective here wasn’t to overturn anyone’s results or call out particular studies. The project was designed to conduct fair and direct replications, Nosek said. “Before we began, we tried to define a protocol to follow so that we could be confident that every replication we did had a fair chance of success.” Before embarking on their studies, replicators contacted the original authors and asked them to share their study designs and materials. Almost all complied.

Researchers who conducted the replication studies also asked the original authors to scrutinize the replication plan and provide feedback, and they registered their protocols in advance, publicly sharing their study designs and analysis strategies. “Most of the original authors were open and receptive,” project coordinator Mallory Kidwell told me.

Despite this careful planning, less than half of the replication studies reproduced the original results. While 97 percent of the original studies produced results with a “statistically significant” p-value of 0.05 or less,5 only 36 percent of the replication studies did the same. The mean effect sizes in the replicated results were less than half those of the original results, and 83 percent of the replicated effects were smaller than the original estimates.

These replication studies can’t explain why any particular finding was not reproduced, but there are three general possibilities, Nosek said. The originally reported result could have been a false positive, the replication attempt may have produced a false negative (failing to find an effect where one does exist), or the original study and the replication could both be correct but arrive at disparate results because of differences in methodology or conditions that weren’t apparent.

The best predictor of replication success, Nosek told me, was the strength of the original evidence, as measured by factors such as the p-value. Yes, the p-value — that notoriously misleading statistic. This study suggests that p-values can provide useful information, Nosek said. “If it was good for nothing, it wouldn’t have shown any predictive value at all for reproducibility.”

At the same time, this project’s results serve as a stark reminder that the 0.05 threshold for p-values is arbitrary. “What it suggests is that when we get a p-value of 0.04, we should be more skeptical than when we get a lower value,” Nosek said.

And, of course, the study provides insights into the prevalence of false positives in psychology research. “I think this is probably about the best estimate that we have at this point,” said Norbert Schwarz, a psychologist at the University of Southern California who was not involved in the project and calls himself a “skeptical observer of the ‘replication movement.’” Schwarz has criticized some previous replication attempts but says this one is “a much more credible and serious attempt than these often haphazard replications driven by a certain vigilante spirit.”

While the results don’t suggest that psychology is a bunch of rubbish, the project does document that psychology has a problem, University of Oxford psychologist Dorothy Bishop told me. “We shouldn’t be sitting around saying who’s to blame, but taking steps to counteract the difficulties.”

Registered reports6 are one approach to reducing reproducibility problems that has been adopted at more than a dozen journals. In a registered study, protocols, methods and data analysis plans are registered and peer-reviewed before data is collected.

“It does make your science much better, because you’re forced to think through your protocol in advance, and you get refereeing and all these useful suggestions before it’s too late,” Bishop said. Committing to methods in advance can also cut down on p-hacking — adjusting the parameters of the analysis until you get a statistically significant p-value — and other intentional or inadvertent post hoc tinkering that may lead to false positives.

This replication project followed a registered reports model. Getting a clear picture of what the original authors did was the biggest hurdle that replicators faced, project coordinator Johanna Cohoon told me, so adopting more transparent methods could make future replication efforts easier.

While some people may look at the high rate of failure to replicate in this study and wonder whether psychology is bunk, the researchers I spoke with took a different view. “Good colleagues that I respect a lot think the sky is falling down, but I don’t agree,” said Elizabeth Gilbert, a University of Virginia graduate student who did one of the studies that failed to replicate the original. Some published results probably are false, she said, but in many cases, the original study may simply have overestimated the effect size, or the phenomenon is found only under particular conditions that haven’t yet been identified. Digging into these problems doesn’t weaken psychology, she says, it strengthens it. “I’m really proud of our field for trying to push forward to make our science better.”

Read more: Science Isn’t Broken: It’s just a hell of a lot harder than we give it credit for

Footnotes

  1. Inclusion of non-live-interview polls does not change this analysis. ^
  2. If a poll tested multiple primary matchups — with Joe Biden and without Biden — I used the question with Biden. ^
  3. His 30 percent in the Des Moines Register poll was just south of the 33 percent he earned in a June Quinnipiac University poll and the 31 percent he took in a CNN poll from earlier this month. ^
  4. Sorry, Jim Gilmore — we didn’t collect data on you. ^
  5. A p-value is simply the probability of getting a result at least as extreme as the one you saw if your hypothesis is false. By convention, a p-value less than or equal to 0.05 is considered “statistically significant” in psychology. ^
  6. Since about 2000, the FDA has required drug trials to register their design and protocols in advance at ClinicalTrials.gov. A study published this month in PLOS One found that the percentage of clinical trials funded by the National Heart, Lung, and Blood Institute that found positive results dropped from 57 percent to 8 percent after the rule took effect. ^

Comments Add Comment

The New York Mets are on some kind of tear right now. With a 9-4 victory over the Philadelphia Phillies on Wednesday night, the Mets have won 21 of their last 29 games — a run that dates to July 25 and the handful of player acquisitions the team made before MLB’s trade deadline.

The biggest catalyst for New York’s sudden success has been a vastly improved offense. Over that red-hot 29-game stretch, they’re averaging 6.2 runs per contest, or 46 percent more than the National League average.7 By contrast, the Mets were scoring just 3.4 runs per game before their hot streak began, or 11 percent below the NL average. New York’s shift from ranking 26th in park-adjusted runs per game8 through July 24 to No. 1 in the month-plus since is by far the biggest offensive turnaround any team enjoyed over that span:

paine-datalab-mets-table

But how much of this improvement should we expect the Mets to retain going forward? To get a general sense of how much regression to the mean tugs on scorching August performances like those of the 2015 Mets, I gathered Retrosheet data on all MLB teams in the expansion era9 and measured how much of a team’s August scoring-index boost (relative to its scoring index through July) carried over into September and October.10 The effect was slight, but significant: About 21 percent of a team’s August hike in scoring rate is “real,” in the sense that it can be expected to continue into subsequent months.

And the Mets have reason to believe they can hang on to more than 21 percent of their offensive gains. One of their trade-deadline pickups, outfielder Yoenis Cespedes, leads the team in offensive runs above average over the past 30 days. So at least part of New York’s scoring surge is due to talent they’ve added since July (as opposed to existing players simply hitting better, which is less sustainable). And, as our friend Jonah Keri pointed out Monday, the Mets also have the easiest schedule in baseball over the remainder of the season.

Combine it with New York’s 86 percent probability of making the playoffs, and the Mets are looking like a far more credible World Series threat than they were earlier in the season.

Footnotes

  1. Inclusion of non-live-interview polls does not change this analysis. ^
  2. If a poll tested multiple primary matchups — with Joe Biden and without Biden — I used the question with Biden. ^
  3. His 30 percent in the Des Moines Register poll was just south of the 33 percent he earned in a June Quinnipiac University poll and the 31 percent he took in a CNN poll from earlier this month. ^
  4. Sorry, Jim Gilmore — we didn’t collect data on you. ^
  5. A p-value is simply the probability of getting a result at least as extreme as the one you saw if your hypothesis is false. By convention, a p-value less than or equal to 0.05 is considered “statistically significant” in psychology. ^
  6. Since about 2000, the FDA has required drug trials to register their design and protocols in advance at ClinicalTrials.gov. A study published this month in PLOS One found that the percentage of clinical trials funded by the National Heart, Lung, and Blood Institute that found positive results dropped from 57 percent to 8 percent after the rule took effect. ^
  7. After adjusting for park effects. ^
  8. Indexed relative to the league. ^
  9. Since 1961. ^
  10. Including regular-season games only. ^

Comments Add Comment

More >

Never miss the best of FiveThirtyEight.

Subscribe to the FiveThirtyEight Newsletter
×

Sign up for our newsletters to keep up with our favorite articles, charts and regressions. We have three on offer: a curated digest of the best of FiveThirtyEight from the past week; The Week In Data, our weekly look at the best data journalism from around the web; and Significant Digits, our roundup of numbers in the news. Enter your email below, and we’ll be in touch.





By clicking subscribe, you agree to the FanBridge Privacy Policy

Powered by WordPress.com VIP