Short Takes: Gerrymandering, Effect Sizes, Race Times and More

I seem to have a lot of articles piling up that I have something to say about, but not enough for a full post. Here’s 4 short takes on 4 current items:

Did You Hear the One About the Hungry Judges?
The AVI sent me an article this week about a hungry judge study I’ve heard referenced multiple times in the context of willpower and food articles. Basically, the study shows that judges rule in favor of prisoners requesting parole 65% of the time at the beginning of the day and 0% of the time right before lunch. The common interpretation is that we are so driven by biological forces that we override our higher order functioning when they’re compromised. The article rounds up some of the criticisms of the paper, and makes a few of its own…namely that an effect size that large could never have gone unnoticed. It’s another good example of “this psychological effect is so subtle we needed research to tease it out, but so large that it noticeably impacts everything we do” type research, and that should always raise an eyebrow. Statistically, the difference in rulings is as profound as the difference between male and female height. The point is, everyone would know this already if it were true. So what happened here? Well,this PNAS paper covers it nicely but here’s the short version: 1) the study was done in Israel  2) This court does parole hearings by prison, 3 prisons a day with a break in between each 3) prisoners who have legal counsel go first 4) lawyers often represent multiple people, and they chose the order of their own cases 5) the original authors lumped “case deferred” and “parole denied” together as one category. So basically the cases are roughly ordered from best to worst up front, and each break starts the process over again. Kinda makes the results look a little less impressive, huh?

On Inter-Country Generalization and Street Harassment
I can’t remember who suggested it, but I saw someone recently suggest that biology or nutrition papers in PubMed or other journal listings should have to include a little icon/picture at the top that indicated what animal the study was done on. They were attempting to combat the whole “Chemical X causes cancer!” hoopla that arises when we’re overdosing mice on something. I would like to suggest we actually do the same thing with countries, maybe use their flags or something. Much like with the study above, I think tipping people off that we can’t make assumptions things are working the same way they work in the US or whatever country you hail from. I was thinking about that when I saw this article from Slate with the headline “Do Women Like Being Sexually Harassed? Men in a New Survey Say Yes“. The survey has some disturbing statistics about how often men admit to harassing or groping women on the street (31-64%) and why they do it (90% say “it’s fun”), but it’s important to note it surveyed men exclusively in the Middle East and Northern Africa. Among the 4 countries, results and attitudes varied quite a bit, making it pretty certain that there’s a lot of cultural variability at play here. While I thought the neutral headline was a little misleading on this point, the author gets some points for illustrating the story with signs (in Arabic) from a street harassment protest in Cairo. I only hope other stories reporting surveys from other countries do the same.

Gerrymandering Update: Independent Commissions May Not be That Great (or Computer Models Need More Validating)
In  my last post about gerrymandering, I mentioned that some computer models showed that independent commissions did a much better job of redrawing districts than state legislatures did. Yet another computer model is disputing this idea, showing that they aren’t. To be honest I didn’t read the working paper here and I’m a little unclear over what they compared to what, but it may lend credibility to the Assistant Village Idiot’s comment that those drawing district maps may be grouping together similar types of people rather than focusing on political party. That’s the sort of thing that humans of all sorts would do naturally and computers would call biased. Clearly we need a few more checks here.

Runner Update: They’re still slow and my treadmill is wrong
As an update to my marathon times post, I recently got sent this websites report that  showed that US runners for all distances are getting slower. They sliced and diced the data a bit and found some interesting patterns: men are slowing down more than women and slower runners are getting even slower. However, even the fastest runners have slowed down about 10% in the last two decades. They pose a few possible reasons: increased obesity in the general population, elite runners avoiding races due to the large numbers of slower runners, or in general leaving to do ultras/trail races/other activities. On a only tangentially related  plus side, I thought I was seriously slowing down in my running until I discovered that my treadmill was incorrectly calibrated to the tune of over 2 min/mile.  Yay for data errors in the right direction.



Born to Run Fact Check: USA Marathon Times

I’ve been playing/listening to a lot of Zombies, Run! lately, and for a little extra inspiration I decided to pull out my copy of “Born to Run” and reread it. Part way through the book I came across a statistic I thought was provocative enough that I decided to investigate it. In a chapter about the history of American distance running, McDougall is talking about the Greater Boston track club and says the following:

“…by the early ’80s, the Greater Boston Track club had half a dozen guys who could run a 2:12 marathon. That’s six guys, in one amateur club, in one city. Twenty years later, you couldn’t find a single 2:12 marathoner anywhere in the country.”

Now this claim seemed incredible to me. Living in Boston, I’d imagine I’m exposed to more marathon talk every year than most people, and I had never heard this. I had assume that like most sports, those who participated in the 70s would be getting trounced by today’s high performance gear/nutrition/coached/sponsored athletes. Marathoning in particular seems like it would have benefited quite a bit from the entry of money in to the sport, given the training time required.

So what happened?

Well, the year 2000 happened, and it got everyone nervous.

First some background In order to make the US Olympic marathon team, you have to do two things 1) finish as one of the top 3 in a one off qualifying race 2) be under the Olympic qualifying time.  In 1984, pro-marathoners were allowed to enter the Olympics. In 1988, the US started offering a cash prize for winning the Olympic trials. Here’s how the men did, starting from 1972:

I got the data from this website and the USATF. I noted a few things on the chart, but it’s worth spelling it out: the winners from 1976 and 1984  would have qualified for every team except 2008 and 2012. The 1980  winner would have qualified for every year except 2012, and that’s before you consider that the course was specifically chosen for speed after the year 2000 disaster.

So it appears to be relatively well supported that the guys who were running marathons for fun in the 70s really would keep pace with the guys today, which is pretty strange. It’s especially weird when you consider how much marathoning has taken off with the general public in that time. The best estimates I could find say that 25,000 people in the US finished a marathon in 1976, and by 2013 that number was up to about 550,000. You would think that would have swept up at least a few extra competitors, but it doesn’t look like it did. All that time and popularity and the winning time was 2 minutes faster for a 26 mile race.

For women it appears to be a slightly different story. Women got their start with marathoning a bit later than men, and as late as 1967 had to dodge race officials when they ran. Women’s marathoning was added to the Olympics in 1984, and here’s how the women did:

A bit more of a dropoff there.

If you’ve read Born to Run, you know that McDougall’s explanation for the failure to improve has two main threads: 1) that shoe companies potentially ruined our ability to run long distances and 2) that running long distances well requires you to have some fun with your running and should be built on community. Both seem plausible given the data, but I wanted to compare it to a different running event to see how it stacked up. I picked the 5000 m run since that’s the most commonly run race length in the US. The history of winning times is here, and the more recent times are here. It turns out the 5k hasn’t changed much either:

So that hasn’t changed much either….but there still wasn’t a year where we couldn’t field a team. Also complicating things is the different race strategies employed by 5000m runners vs marathon runners. To qualify for the 5k, you run the race twice in a matter of a few days. It is plausible that 5k runners don’t run faster than they have to in order to qualify. Marathon runners on the other hand may only run a few per year, especially at the Olympic level. They are more likely to go all out. Supporting this theory is how the runners do when they get to the Olympics. The last man to win a 5000m Olympic medal for the US is Paul Chelimo. He qualified with a 13:35 time, then ran a 13:03 in the Olympics for the silver medal. Ryan Hall on the other hand (the only American to ever run a sub 2:05 marathon), set the Olympic trials record in 2008 running a 2:09 marathon. He placed 10th in the Olympics with a 2:12.  Galen Rupp won the bronze in Rio in 2016 with a time 1 minute faster than his qualifying time. I doubt that’s an unusual pattern….you have far more control over your time when you’re running 3 miles than when you’re running 26.  To further parse it, I decided to pull the data from the Association of Road Racing Statisticians website and get ALL men from the US who had run a sub 2:12 marathon. Since McDougall’s original claim was that there were none to be found around the year 2000, I figured I’d see if this was true. Here’s the graph:

So he was exaggerating. There were 5.

Pedantry aside, there was a remarkable lack of good marathoners in those years, though it appeared the pendulum started to swing back. McDougall’s book came out in 2009 and was credited with a huge resurgence in interest in distance racing, so he may have partially caused that 2010-2014 spike. Regardless, it does not appear that Americans have recaptured whatever happened in the early 80s, even with the increase in nearly every resource that you would think would be helpful. Interestingly enough, two of the most dominate marathoners in the post-2000 spike (Khalid Khannouchi and Meb Keflezighi) came here in poverty as immigrants when they were 29 and 12, respectively. Between the two of them they are actually responsible for almost a third of the sub-2:12 marathons times posted between 2000 and 2015. It seems resources simply don’t help marathon times that much. Genetics may play a part, but it doesn’t explain why the US had such a drop off. As McDougall puts it “this isn’t about why other people got faster; it’s about why we got slower.”

So there may be something to McDougall’s theory, or there may be something about US running in general. It may be that money in other sports siphoned off potential runners, or it may be that our shoes screwed us or that camaraderie and love of the sport was more important than you’d think. Good runners may run fewer races these days, just out of fear that they’ll get injured. I don’t really know enough about it, but the stagnation is a little striking. It does look like there was a bit of an uptick after the year 2000 disaster….I suspect seeing the lack of good marathon runners encouraged a few who may have focused on other sports to dive in.

As an interesting data point for the camaraderie/community influence point, I did discover that women can no longer set a marathon world record in a race where men also run.  From what I can tell, the governing bodies decided that being able to run with a faster field/pace yourself with men was such an advantage that it didn’t count. The difference is pretty stark (2:15 vs 2:17), so they may have a point. The year Paula Radcliffe set the 2:15 record in London, she was 16th overall and presumably had plenty of people to pace herself with. Marathoning does appear to be a sport where your competition is particularly important in driving you forward.

My one and only marathon experience biases me in this direction. In 2009 I ran the Cape Cod Marathon and finished second to last. At mile 18 or so, I had broken out in a rash from the unusually hot October sun, had burst in to tears and was ready to quit. It was at that moment that I came across another runner, also in tears due to a sore knee. We struck up a conversation and laughed/talked/yelled/cried at each other for the remaining 7 miles to the finish line. Despite my lack of bragging rights for my time I was overjoyed to have finished, especially when I realized over 400 people (a third of entrants)  had dropped out. I know for a fact I would not have made it if I hadn’t run in to my new best friend at that moment of despair, and she readily admitted the same thing. McDougall makes the point that this type of companionship running is probably how our ancestors ran, though for things like food and safety as opposed to a shiny medal with the Dunkin Donuts logo. Does this sort of thing make a difference at the Olympic level? Who knows, but the data and anecdote does suggest there’s some interesting psychological stuff going on when you get to certain distances.

Race on folks, race on.