Do Soccer Players Choke?
Assessing the research on penalty kicks and pressure.
While Spain is a modest favorite over the Netherlands to win Sunday's World Cup final, the matchup is not lopsided by any means. That means there's a reasonably good chance that, after 120 minutes of play between two fairly even teams, the game will come down to penalty kicks. Whether you greet that prospect with a casual fan's delight or a purist's horror, there is no question that a penalty shootout is dramatic and stressful. If the game does go to penalties, there will be at least one moment when the entire tournament—and with it a lifetime of glory or infamy—comes down to a single kick, just as last week's infamous Ghana-Uruguay came down to a missed penalty by Ghanaian striker Asamoah Gyan.
How would you hold up under that kind of pressure? That's an ongoing question in sports psychology—whether people sometimes exhibit a tendency to "choke" when they're under pressure or, alternatively, to raise their level of performance when the situation demands it. According to a handful of published studies, choking does exist in soccer. Nevertheless, I'm a choking skeptic—from where I sit, the evidence that penalty kickers succumb to pressure is very weak.
The "Yerkes-Dodson Law" predicts that participants in a penalty shootout should buckle under pressure. According to the theory, human performance follows an "inverted U shape." Under the effect of mild stress, or "arousal," proficiency improves as the subject expends more concentration and energy. But past a certain point, too much pressure leads to panic and attention problems, and choking ensues.
Outside the laboratory, it's hard to come up with experiments to test the "inverted U" theory. How can you determine how much stress a subject is under? And how can you find a group of subjects who do exactly the same task, so you'll be able to compare them properly?
That's where penalty kicks come back in. Psychologists love to study penalties because they provide easy answers to the above questions. How much stress is the subject under? That can be estimated from the score and the game situation. Plus, every penalty kick is taken under exactly the same rules and conditions (save the identity of the goalkeeper, which, the consensus is, doesn't matter nearly as much as the identity of the kicker).
The penalty kick, then, seems like the perfect laboratory to study how we respond to pressure. A handful of researchers, after sifting through penalty shootout data, have determined that soccer players do show a tendency to choke under stress. In one study, described in last month's New York Times, researcher Gier Jordet looked at a dataset of important international matches and found that when a team needed an immediate goal to win a shootout (and the game), it succeeded 90 percent of the time. But when a team needed a goal simply to tie the shootout, and a miss would mean an immediate loss, it succeeded only 60 percent of the time. (The overall average on penalties was 79 percent, as Jordet and colleagues reported in this paper.) David A. Savage and Benno Torgler found similar results in their own study. (The Savage-Torgler paper doesn't seem to give a nontechnical interpretation of the numbers, but if I understand their method properly, their findings seem comparable to Jordet's.)
Still, despite the two studies reaching such similar conclusions, I'm not convinced that penalty kickers choke under pressure.
Why not? First, the effect—a 90 percent success rate on "immediate win" shots compared with 60 percent on "immediate loss" shots—seems implausibly large. Implicit in these studies is the assumption that all players respond the same to stress. But isn't it reasonable to assume that athletes react to stress differently? For any given stressful situation, some of the kickers might be in the rising portion of the "inverted U" and still be at their best. Suppose, then, that in an "immediate loss" situation, half the kickers are still "clutch," with 90 percent conversion rates, and the other half choke. For the average of the two groups to wind up at the observed rate of 60 percent, the "chokers" would have to convert at the rate of 30 percent. That just doesn't seem right. Not only is that figure absurdly low, but you also have to accept that shooting for the win is much, much less stressful than shooting for the tie—so much less stressful, in fact, that it bumps your chances all the way up from 30 percent to 90 percent. Does that really seem reasonable?
Second, we need to think about the context of the "immediate loss" kicks. In a penalty shootout, each team gets to pick its five best shooters. If the contest is still tied after five kicks apiece, the shootout goes to "sudden death"—if the first team makes its kick, the second team has to score or it will lose the game. In this sudden death phase, each squad must deploy its lesser shooters, the preferred five having already been used in the first phase of the shootout. It's in these situations—when less-skilled players are likely to be shooting—that many of the "immediate loss" kicks happen.
Suppose the score is tied after five kicks each. The first team's sixth kick will never be an "immediate win" or an "immediate loss" kick—it either takes the lead or preserves the tie. The second team's sixth kick, on the other hand, will always be an "immediate win" or "immediate loss" kick. But it will be an "immediate loss" kick much more often than an "immediate win" kick, because the first team will have scored on its previous kick much more often than it will have missed. The "immediate loss" category, then, will almost certainly carry a larger proportion of worse players than the "immediate win" group. That means we should expect worse performance on "immediate loss" kicks even if stress plays no role at all.
How significant is this factor? I don't really know for sure. My gut feeling is that it's big enough to explain the statistical significance of the small "immediate win" improvement, but not enough to negate the larger "immediate loss" difference the studies found.
It's also worth noting that even though the results in these reports were reported as statistically significant, they're based on fairly small samples. The Savage/Torgler study was based on only 49 "immediate loss" kicks, meaning the "choking" conclusion is based on a mere handful of extra misses.
In search of a larger sample, I was led by Google Scholar to this study by Thomas J. Dohmen. Dohmen, who looked at the 40-year history of in-game penalty kicks in Germany's Bundesliga, found one situation that was statistically significant: home kickers were more likely to "miss by choking." (Dohmen's definition of choking: missing the net or hitting the posts or crossbar.) The suggestion here is that the expectations of the home crowd stress the kicker to the extent that he's more likely to miss completely.
The finding was based on 3,619 penalty kicks, about 10 times as many as the other studies. Still, the narrow definition of "choke" meant that there were only 59 such misses by the visiting teams. If you took 15 chokes by home teams and handed them to visiting teams—that's one choke every three years—the effect would disappear. And, of course, there's the more obvious criticism: If playing at home is so stressful, how do you explain home field advantage?
So: Three studies, all of which claim to have confirmation that penalty kickers choke. Still, I'm skeptical that Asamoah Gyan choked when he missed that kick for Ghana. And if some poor Spanish or Dutch player botches a penalty kick on Sunday, I won't be calling him a choker either. If you miss a kick in the World Cup final, though, I'm not sure that's much of a comfort.