Management 101: It really does work

Do those management practices make sense as ways to improve an organization's performance? Many of us probably think so, but in truth, most of us don't really know anything about whether such practices really work beyond instinct, anecdote and preconception.

This is why God invented scholarly research. Now three public administration academics, Nathan Favero and Ken Meier of Texas A&M and Larry O'Toole of the University of Georgia, have published a fascinating paper on such questions. It carries the straightforward title of Goals, Trust, Participation, and Feedback: Linking Internal Management with Performance Outcomes, and appears in the new issue of the Journal of Public Administration Research and Theory, the leading scholarly journal on public management.

There are two challenges to doing research that can shed light on whether following these practices in fact improves performance. The first is gathering information on an organization's performance, and hopefully similar information from a number of organizations so their performance can be compared quantitatively. (One reason there is more research on the impact of management practices on business firms than government organizations is that getting data to capture and compare performance of companies -- such as stock price or return on assets -- is far easier than for government.) The other challenge is getting information about what practices organizational leaders actually use.

If you can do both of those things, then your strategy as a researcher is fairly straightforward. You compare the performance of organizations that use a practice with those that don't, and see if there are any patterns of difference. (You need to hold constant, or in research jargon control for, differences between the more and less successful that may be correlated with success, but not caused by different management practices. So, for example, schools with privileged students are likely to perform better than those with disadvantaged kids, independently of whether the schools follow the management practices being investigated.)

The first challenge is easier in the public sector if there are organizations with a lot of units delivering the same service. Two prominent examples are schools and hospitals. And, before this paper came out, there was already a cottage industry of scholars (including Meier himself) analyzing data on schools' performance on standardized tests. Of course this isn't a perfect measure, but compared to many others, it's not bad at all.

The bigger problem is measuring use of these various management practices. Sending in outside observers to look is incredibly costly, and precludes analysis of any large number of cases. Asking managers to self-report their behavior is plagued by the grade inflation and distortion to which self-reports are prone.

Favero and colleagues came up with a great idea: they gathered information on principals' behaviors by using an existing survey that asked not the managers but the people being managed -- their subordinates (teachers) -- about them. One of several questions to measure goalsetting, for example, was, "My school has clear measures of progress for student achievement throughout the year." For feedback, it was, "School leaders visit classrooms to observe the quality of teaching at this school."

The data for the paper all come from New York City schools. And what did the researchers find? When you put all four Management 101 practices together, and added the various controls, three of the four practices had a noticeable impact on improving student test scores.

By a considerable margin (the researchers were able to compare the relative impact of one practice compared to another on test scores), the most helpful of the four practices was goal setting. Furthermore, the impact of the other practices was strengthened when done in conjunction with goal-setting. Chalk one up for performance measurement.

Guess which practice actually had a negative impact on test scores? I suspect it will surprise many readers that it was trust/commitment building, measured by questions such as "School leaders communicate a clear vision for the school" and "I trust the principal at his/her word." Interestingly, this was similar to a finding I had in research on the impact of management practices in a totally different context, anti-crime partnerships in British local government. Hard to know why, but perhaps spending time on this didn't accomplish anything, but took time away from more useful practices.

There's more. The authors also looked at measures of school success other than test scores -- such as attendance, parental satisfaction with the school, and even school violence (they were able to get data for all these) -- and likewise found that the Management 101 practices were also related, though more weakly, to success on those dimensions.

Many readers of this blog of course work for organizations whose performance cannot be measured as well as schools, especially in a way that allows comparisons across a large number of units. Yet these findings are still relevant to you. Of course, not all organizations are the same, and we can't be certain that techniques that work well in schools would work in your organization. But, in the absence of contrary information, it is not a bad bet. We should all thank these three scholars for their work. All you federal managers out there, let's move from knowledge to action.