Do High Fives improve performance?
One side-story from the recent NBA Finals involved the link between team performance and high fives, chest pats, and other positive physical contact between players. It even made the Wall Street Journal after game 4 (“Dallas’s Secret Weapon: High Fives”, June 9, 2011).
The story was based on recent research from Michael Kraus, Cassy Huang, and Dacher Keltner out of the University of California, Berkeley, titled “Tactile Communication, Cooperation, and Performance: An Ethological Study of the NBA.” The authors concluded physical touch positively impacted player and team performance, even after controlling for player salaries and a team’s pre-season expectations. To their credit, the authors acknowledged that other factors, such as leadership style or personality, may have accounted for the observed relationship between touch and performance. But this caveat is largely buried in the text.
Although the study makes for fun headlines and sports banter, it also makes one of the more common mistakes in statistics and econometrics – confusing correlation with causation. On average, teams that perform well early in the season will tend to also perform well at the end of the season. There are certainly counter-examples here, but regressions only model the average, so we’ll restrict our logic accordingly. In addition, teams that perform well are likely to have more camaraderie and generally more events worth celebrating – thus, more high fives and celebratory touching. The observed relationship between early-season touching and late season performance therefore seems derived from the relationship between early season performance and late season performance. Unfortunately, this result (that early season performance predicts late season performance) is far less sexy.
This is not a criticism of the study’s goal or ultimate interest; rather, it is a criticism of the media’s interpretation and, to a lesser extent, the authors’ methodology. Revealing interesting correlations in data is a worthwhile pursuit, and this study clearly succeeds in that area, but interpreting correlation with causation (without the appropriate methodological adjustments) continues to be an all-too-common error in statistical analysis and public interpretation.
Comment on this article
Skip to Main Site Navigation / Login
Comments
Ellie
I feel like I must be missing something. Like it's described in the post above, of course there's more positive physical interaction between members of teams who are performing well, both individually and as a group. You slam dunk, you get head slaps and bear hugs; you all work together to make an amazing play, you celebrate your collective awesomeness.
But that being said, what would be interesting to see is if, among teams who performed well, if an increase in physical touching was a good indicator of future performance. Because teams can have strong players and success, but they don't always "gel". Look at only winners or teams that made it to the playoffs or group individuals and teams by points scored or plays made, etc., and look at short (game-length) and long term (full season) effects. Or even across seasons...maybe the introduction of one player to a team who has shown to be more tactile - touchy-feeley- than other players has an impact on the success of the team when matched with players of similar ability who are less physically social...
reply to this comment