Steve Kelman reports that better data is producing a mini-revolution in scholarship about public-sector management.
The holy grail of scholarly research on management is to find credible evidence that certain management or leadership practices actually improve an organization’s performance.
The key word here is “credible.” For years, finding this holy grail has been inhibited by methods used for management research that challenge the credibility of any purported findings. There have traditionally been several problems with the data on which such research is based. Measurements of what kinds of practices organizations use have usually been based on surveys where organization members self-report what practices are used. Measurements of performance have similarly been based on survey self-reports of how well respondents feel their organization is doing.
This approach is very problematic. Data both on the organization’s practices and on its performance are gathered from the same survey respondents. This creates a real tendency to dramatically inflate associations between practices and performance.
The underlying cause of responses, both about practices and performance, is often the respondent’s overall attitude to the organization. So if they like their organization, they will respond that the organization uses various practices that tend to be seen positively, and also to report that their organization performs well. But there is not enough separation between these two kinds of reports; reporting one thing is associated with reporting another, independent of any real causal influence of the practices on performance.
This worry is exacerbated further when the measure of performance is also the respondent’s subjective judgment of performance. Lots of the research based on the Federal Employee Viewpoint Survey has these issues.
In recent years, scholars have tried to mitigate these problems by gathering data on practices from a different source than the one used to gather data on performance, and by using objective rather than subjective performance data. This is producing a mini-revolution in scholarship about public-sector management.
In a beautiful paper that recently went online in the Public Administration Review, Rusi Sun of the University of Michigan at Dearborn and Alexander Henderson of Long Island University exemplify the improvements that have been taking place in management scholarship. They look at whether a leadership style by school principals called “transformational leadership” is associated with higher standardized test scores for students in those schools.
“Leadership,” of course, means lots of different things to different people, so it is important to note the features of the kind of leadership the paper discusses. “Transformational leadership” in the academic leadership literature refers to a style that goes beyond carrots and sticks as ways to motivate employees.
Transformational leaders motivate organization members to look beyond their self-interest to consider organizational goals. They provide an appealing vision that generates enthusiasm and strengthens commitment to the organization’s goals, as well as promoting employees' intellectual stimulation.
The authors note that typically the literature on transformational leadership emphasizes one-on-one relationships between leaders and followers. However, their research examines an alternative way transformational leaders seek to improve the organization’s performance – by putting in place organizational practices and processes that encourage improved performance. These efforts focus not just on the one-on-one retail level, but on the wholesale level of changing how the organization as a whole works. The three organizational practices the paper associates with transformational leadership are using performance information to manage the organization, engaging external stakeholders (parents) in the work of the school, and building a collaborative culture inside the school.
Note how the authors have gathered their data. Their measure of the organization’s performance is not a subjective report by teachers or principals but the average student scores on standardized tests, controlling for differences across schools (such as the socio-economic status of the parents) that have an impact on scores. They measure the extent to which the principal has a transformational leadership style (such as promoting an appealing vision) through a survey of teachers -- the leaders’ subordinates -- rather than self-reports of the principals. They measure the specific transformational leadership practices (using performance information, and promoting stakeholder engagement and a collaborative culture) in different ways. To measure stakeholder engagement, they use a survey of parents on the extent to which principals communicate with parents and involve them in meetings, not the self-reports of principals. To measure use of performance information and collaborative culture, they survey teachers rather than principals.
The authors argue that transformational leaders are more likely to promote these practices because they communicate an inspiring vision for subordinates that looks beyond self-interest and strengthens commitment to the organization’s goals. The practice of using performance information can improve performance by inspiring teachers to be committed to organizational improvement and learning, and to accepting that it is worthwhile to devote time to using information for improvement efforts. The practice of parental engagement can promote better student performance by signaling to teachers and parents alike that the mission of the organization is important, and that parental participation can help in achieving the mission.
So what do Sun and Henderson find? A principal’s transformational leadership was associated with higher standardized test scores, though only to the extent that the principals translated their leadership style into the organizational practices of using performance information and involving parents. (They did not find that promoting a collaborative culture among teachers aided student performance.)
These results control, as a good scholarly study should, for various variables that might influence the results, such as the socio-economic status of the parents. (Interestingly, the study found that schools where the parents had lower socio-economic status had higher parental engagement in the school).
The bottom line: Leadership styles can matter for an organization’s performance. The next step for researchers – to give findings such as these more practical bite -- is to ask whether it is possible for public managers to be trained to improve their leadership styles. Interestingly, a group of Danish researchers is studying just this question in the context of the public sector in Denmark. In a subsequent blog, I will look at what they are learning.
It is nice to see that at least some academics studying public management are now asking practical questions in academically rigorous ways.