the lectern banner

By Steve Kelman

Blog archive

This could be the start of something big

Shutterstock image.

When it comes to future performance, how important are the early managerial choices made by leaders of a new organization -- or by leaders of a new cross-organizational collaborative enterprise or new unit within an organization?

I would guess that many of us suspect these early choices are often important. And there is a whole strand of social science research, called "path dependence theory," that argues that people, organizations, and even countries can early on start on a path, maybe even by chance, that gets reinforced over time: The classic example from the literature is the QWERTY keyboard system, which was originally developed to slow people down using a typewriter because the keys would jam if pounded too quickly, but remained in the age of electric typewriters and then computers because people had invested in learning the old system.

I often discuss the research of other scholars in this blog, and, prompted by the spread of online transmission of academic papers, I have now decided briefly to discuss some research of mine (together with Sounman Hong of Yonsei University in Korea), which has just been published online at the Journal of Public Administration Research and Theory.

I have been studying a cross-agency anti-crime collaboration in British local government called Crime and Disorder Partnerships, whose mission has been to promote collaboration across agencies to fight crime. Established in 1998 in local governments throughout England, these partnerships involve the police, probation, youth social workers, the fire department and local city service organizations such as lighting.

It is possible to measure their success by looking at differences across the local partnerships in changes in crime rates over the following decade. I undertook a survey of the relatively modest number of first leaders of these partnerships I was able to locate, and asked them a bunch of questions, including asking them to rank the importance of different management priorities that could be used to get the organization off to a good start. Those priorities ranged from working to create trust among partners to promoting a common vision for the partnership to establishing a performance measurement system and following up as a leader to make sure partners kept the commitments they made.

What I found in the paper is that these early choices actually did have an impact on crime a decade later! Partnerships whose leaders got them off to a good start through some kinds of choices had better crime performance than those who made different choices.

So what kinds of early choices worked and what kinds didn't?

This will obviously depend on the type of organization and the situation they face. But these partnerships needed to bring about an organizational change in the participating organizations, which were not used to collaboration and often had mutually suspicious cultures (for example, police and social workers).

I divided up the management approaches into those that tried to initiate change by changing first the attitudes of participants (e.g. trust building) versus those that tried to start by changing behaviors (e.g. getting partners to share information). The results showed that prioritizing measures to change attitudes first did not reduce crime a decade later, while choices that prioritized changing behavior first did improve crime performance.

That's a lesson with general significance for getting organizational change started.

Posted by Steve Kelman on Jun 26, 2014 at 7:51 AM


Featured

Reader comments

Fri, Jun 27, 2014

I was doing something like this 40 years ago: my job was to increase information between Police Forces and councils, fire departments and the like; overcoming the parochial attitudes; sharing information - even if seemingly irrelevant (and it turned out that sooner or later it could be relevant) There was much more to it than I can explain here, but certainly the key is to encourage cooperation on all levels; and to include informal social contacts and meetings and events. We organised pleasure coach trips, sports competitions, so on and so forth. Perhaps a revival is taking place...

Thu, Jun 26, 2014 Steve Kelman

Al, good question. In the statistical model, we "control" for possible influence of age, poverty, London v. non-London, and a bunch of other variables. Thanks for asking!

Thu, Jun 26, 2014 Al

How do you isolate the effects of these decisions on the overall crime rate? Are there any confounding variables you have to control for? Gentrification? Population change? Average Age? Not a challenge to your conclusion, just curious how these things are done . . .

Please post your comments here. Comments are moderated, so they may not appear immediately after submitting. We will not post comments that we consider abusive or off-topic.

Please type the letters/numbers you see above