Click for Academy of Management home page

Academy of Management

Failure teaches best when blame is hard to pin down, study finds

August 26, 2015

For more information, contact: Ben Haimowitz , (718) 398-7642,

Which company problem is more likely to get fixed – one attributable to a particular person or department or one with wider roots? Common sense would suggest the former, since the solution in that case may involve nothing more complicated than changing the way a single unit is run or setting an individual or a few people straight.

But some new research casts doubt on this reasoning, obvious though it may seem. 

In the words of a paper in the current issue of the Academy of Management Journal,"concentrated failures prompt narrower attributions of responsibility which, whether accurate or not, ultimately lead to less thorough investigations and fewer of the system wide changes that are typically required to address organizational performance problems.” 

The study goes on to explain that, “following multiple failures at the same location or involving the same individuals, organizational leaders have a well-recognized tendency to merely dismiss culpable parties or make other relatively localized changes. While these simple changes may be sufficient in some cases, the present study suggests the need for deeper exploration." 

The paper, by Vinit Desai, an associate professor of management at the University of Colorado Denver, comes to these conclusions through an analysis of failures that are literally a matter of life and death – the improvement or lack of it in death rates resulting from cardiac-bypass surgery in more than 115 hospitals in California, where the state gathers copious data on these surgeries. The hospitals that made the greatest strides were those in which deaths were dispersed among patients of a number of surgeons rather than being concentrated in patients of one or just a few. The chance that the degree of concentration did not influence death rates was less than 5%.  

Asked for specific examples, Prof. Desai hesitates to draw conclusions about the results of any single institution as distinct from the sample as a whole. With this caveat, he cites the case of a hospital where over a two-year period about half the surgeons lost patients who underwent this procedure, called Coronary Artery Bypass Graft (CABG). According to the hospital, it responded with broad procedural changes that involved "[putting] evidence-based protocols in place, [providing] educational and skills-based training for staff, and [forming] multidisciplinary committees to identify and resolve quality issues." The following year, no patients died following CABG at the hospital, even though about the usual number of procedures was carried out.  

What about hospitals where bypass deaths were concentrated among a small minority of surgeons? How the departure of subpar physicians affected such institutions individually would be difficult to tell, the professor says. But outcomes were quite clear for the study sample as a whole: greater concentration of failures was associated with more subsequent departures but not with reduced CABG deaths. In short, the exit of less competent surgeons did not, in general, solve the problem. 

Research, the study notes, “has identified a pervasive tendency to overweight individual or dispositional factors and underweight situational ones…When repeated failures primarily involve the same group or employee, organizational decision-makers might suspect that the involved party played a causal or at least a contributory role. This may initially seem reasonable, given the frequency of failures involving this party and the relative rarity of failures involving others within the organization. However, work on defensive attributions would suggest the tendency for decision-makers in this situation to ignore or undervalue any contributing situational factors, and even to overweight the importance of individual or dispositional characteristics.”  

The paper is based on the performance of 116 to 118 hospitals (the number varies over the length of the study) in two-year segments of the period 2003 through 2010. A mean of about five surgeons per hospital, but as many as 10, performed CABG operations, with the doctors averaging about 63 such procedures per year. A little less than three percent of patients died during or after surgery. 

Desai analyzed the relationship between failure concentration (how failure was apportioned among surgeons) in one period, say 2003-04, and the change in the incidence of CABG deaths between those two years and the following two years, say 2005-06. In all, the analysis embraced 468 hospital-period observations, with findings including the following:  

■   The more concentrated failure was in one period, the greater the failure rate in the following period.


 This relationship was stronger in larger hospitals, where concentrations of poor performers stood out more than in smaller places. It was weaker in institutions with CABG death rates either well above or well below the average, the former group being highly motivated to improve their results and the latter group to perpetuate their superiority.


■  Concentrated failure was associated, too, with increased incidence of surgeon departures and fewer hospital efforts to refine CABG procedures.

Asked to sum up the study’s principal lesson, Prof. Desai says, “Don’t be too quick to assign blame. In all likelihood, there is more to be lost in doing so than there is in looking for a deeper cause. Difficult though they may be, deep searches for underlying structural solutions or procedural changes can pay off in a big way even in high-functioning organizations.” 

Indeed, in a 2011 study of the global launch-vehicle industry published in the Academy of Management Journal, Prof. Desai and Peter Madsen of Brigham Young University had occasion to cite a catastrophic instance of this – the tragic loss of seven astronauts aboard the Columbia space shuttle from a structural flaw that had been noticed in a previous shuttle liftoff but went uncorrected because the earlier flight was successful. Organizations, the study concluded, learn more effectively from failure than from success. Now the new research takes the matter a step further, identifying the circumstances in which failure teaches most effectively. 

The paper, “Learning through the Distribution of Failures within an Organization: Evidence from Heart Bypass Surgery Performance," is in the August/September issue of the Academy of Management Journal. This peer-reviewed publication is published every other month by the Academy, which, with close to 20,000 members in 115 countries, is the largest organization in the world devoted to management research and teaching. The Academy's other publications are Academy of Management Review, Academy of Management Perspectives, Academy of Management Learning and Education, Academy of Management Annals, and Academy of Management Discoveries.  

Academy of Management
Member Services
Academy of Management
Online Opportunities
Academy of Management