Wednesday, February 13, 2008

Lessons from Donald Rumsfeld

I confess to have developed a new addiction. At 6:30 or 7:00 CST I switch on CNN and watch whatever nonsense they have on about the election. Because the Superbowl has passed and I’ve lost interest in College Basketball as the Illini are in the second division in the Big Ten and likely won’t finish with a .500 record so won’t make it to a post-season tournament, the Presidential race has become my sporting event to watch. Whether it is speculation about the upcoming primaries, demographic breakdowns of recent primary results, or clips of the candidates addressing their supporters, it all plays out just like a race for a sports championship and creates the same sense of fan interest for me. In this regard I believe I’m typical. I think CNN knows this and delivers its coverage accordingly. For example, they do a lot on age, gender, income, and ethnicity of the voters and which candidates they choose. They’ve done extremely little on issues and, for example, whether a voter prefers Clinton over Obama because of differences in their health care plans. In effect, the candidates themselves have become the issues. That’s how it is with celebrities. That’s what the candidates have become.

Part of the reason to have a primary season with a sequence of primaries (or caucuses) rather than a single primary day on which all voters select the candidates is that the former allows information flows and more full vetting of the candidates so that voters in later primaries make their choices with a greater amount of information than those in the very early ones, Iowa or New Hampshire. The debates have been reasonably good in giving candidates view on the issues as well as getting the voters to understand how the candidates perform public speaking-wise under the pressure and fatigue of the campaign. These news shows, in contrast, seem to be more about us and our reaction to candidates than about the candidates themselves. If a large fraction of such and such group votes for candidate x, then other members of that group in states that haven’t yet held their primaries should also vote for candidate x, to get on the bandwagon. When they talk about a candidate “gaining momentum” it seems to be this bandwagon effect to which they are referring. So my little surmisal of what has been happening this primary season is some vetting of the candidates and some finding of the bandwagon with more of the former early on and more of the latter more recently. Howard Dean will look like a genius if this process produces a clear cut winner before the Democratic convention, without relying on substantial votes from the superdelegates to determine the outcome. But he will look like an absolute ditz if the process deadlocks, because of the disqualification of both Michigan and Florida in selecting delegates for pushing up their primary dates too early, a prize in combination almost as large as Texas and Ohio, the two states Hillary Clinton is counting on to put her over the top.

In Dean’s view, the vetting part is best done in smaller states, where the voters can get closer to the candidates, and having a handful of these with demographically diverse populations should help to give the rest of the voters a good read on the candidates. This time around, with Iowa and New Hampshire followed by Nevada and South Carolina before Super Tuesday, that part worked like a charm, but it really was a two horse race from the start, the type of race where the Dean approach makes the most sense. In the alternative case when there is a prohibitive favorite early on, the whole thing can be sewn up early as has happened in the Republican race, their winner take all approach to allocating delegates making that outcome more likely. But then states like Michigan and Florida are right to want to move their primaries up earlier so they can matter and get the issues of their voters to be accounted for in the political calculus. So the Dean approach works well in the first case but not in the second. I don’t believe there is a single right answer to this, but I hope there is a winner and not deadlock this time around, because we need a different party in the White House and because the Dean approach needs some fine tuning before the up or out decision on it is made.

* * * * *

This recent fascination with politics has got me to be thinking about what parallels there are between politics and learning technology. This is not just idle thought. Sometimes the issues we confront are so contentious that we can’t argue about them or reason through them publicly. Finding parallels in a different field that we are all aware of has value because then we can make the argument about the parallel field without feeling encumbered and then by drawing some conclusions about our own field make some progress in the process. Of course, the parallel world will be dissimilar in many ways and so one must be careful not to draw sweeping conclusions. Please bear that in mind as you read the following.

The current version of the New York Times Magazine has a featured piece on Robert Gates, the Secretary of Defense. The piece is largely positive on what Gates has accomplished since he took office a little more than a year ago. He seems to be well liked both by soldiers in the Armed Forces and by members of the Senate Armed Services Committee, including Carl Levin, the ranking member and a leading Democrat. Much of the article is about how Gates goes about getting things done. Gates is consultative by nature and yet he knows how to move things through a bureaucracy. He is, in some sense, the consummate manager. He is not so much an agenda setter as a relief pitcher, finishing up a game that others started. He is very effective in that role.

Much of Gates success can be attributed to whom he replaced, Donald Rumsfeld, and the contrasting styles of the two. Rumsfeld was an innovator, brash and outspoken and often at odds with the generals who reported to him as well as with the Democrats in Congress. (On the latter the Gates piece in the Times makes clear that some of the behavior is dictated by circumstance. When Rumsfeld was Defense Secretary under Bush, the Republicans controlled both Congress and the White House. Consequently, there was less of a need for consensus building to get things accomplished.) I think it fair to say that many of Rumsfeld’s innovations either failed outright or succeeded in part but did not live up to prior billing. We often talk about learning from our own failures. This blog post is an attempt to learn from the failures of another.

Let me draw out the parallels first. A good part of IT on my campus is provided by a large, bureaucratic organization. My sense is that at other peer institutions the IT organization is likewise large and bureaucratic. Readers who are at smaller institutions can make their own judgments as to whether this particular parallel makes sense on their own campus. The Department of Defense is one of the larger bureaucracies in the country. Both the IT organization and the Department of the Defense provide (congestible) public goods to their respective constituencies. To an economist like me, the primary characterization of a public good is that it is non-excludable, meaning one person’s consumption of the good doesn’t block another person’s consumption of the same good, while congestible means that if the collective consumption is too great then the consumption benefit for any one individual is degraded.

With national defense, there are many possible threats, large and small, and hence a diversity of needs to be satisfied, some which are mutually supporting, other which compete with one another for scarce resources. Likewise with IT and here I’ll just focus on learning technology. The technology can support collaboration between students and instructor and between the students themselves, it can make information more readily available and thereby promote access, it can track behavior and perhaps in this way help in measuring student performance, and it can help students visualize issues that they’d otherwise find difficult to consider. I don’t mean this to be an exhaustive list. I simply want to make the point that because there these competing needs, and because they are addressed by public good provision which in itself doesn’t get subjected to a market test, there can be disagreements about how these needs should be prioritized and which should be addressed first.

A third source of parallel is the view that the world has changed and hence we need to change to accommodate those external changes. At the outset of the Bush II presidency, a widely shared view about the Armed Forces was that they were still largely configured to fight the Cold War and hadn’t yet adjusted to a different set of threats such as the experience in Bosnia-Herzegovina, the specter of a nuclear North Korea, and threats from a variety of other states and groups. These new threats required a flexible and rapid response, one that is tuned to the particular circumstance. Now consider this juxtaposition. The campus IT organization here where I was employed until 15 months ago is now going through a re-org and one of the prime reasons is that it is perceived as too slow to respond to new IT needs as they arise.

The call for change, however, is certainly not restricted to my campus. George Siemens in a very recent paper offers up different possible roles that instructors might play in light of recent IT developments and how those affect the ways students learn. (I found this paper less than fully illuminating because it was not clear which type of instructor should embrace any one of the particular roles that George mentions and further there was not enough drill down on the roles themselves that instructors might make an informed selection themselves based on their own inclinations.) In a similar vein Gardner Campbell, who has been blogging profusely since the ELI conference concluded, wrote a post on Incrementalism, something to be avoided assiduously, citing none other than Brian Hawkins as the authority on the matter, and feeling compelled to introduce his students to Croquet, to get a new vista on collaboration with documents.

For me personally, I’m bipolar regarding change. When it is my own efforts to direct, I’m drawn toward taking a novel, non-orthodox approach. (Who else is talking up Rumsfeld these days, for instance?) But when it comes to inducing change in others, while I may not act as cock-sure of myself as Rumsfeld did and I may offer up my views as suggestions rather than as commandments, I’m afraid I’m guilty of many of his errors, translated back to campus learning technology, of course. So with that, let’s list some of these and discuss the list at least to the extent to explain why they happened. I’m not able to offer up preferred alternatives but I can bring to bear what is at issue from the decision maker’s point of view. Here is the list of issues:

1. Blame those who embody the old way of doing things.

2. Feel pressed in the need to get things done quickly.

3. Let considerations about cost affect the view of what is feasible.

4. Insist that your view of the world is right and hence don’t negotiate to a mutually agreed on position with others who hold contrary opinions.

5. Provide an entry level reason for doing something that is at odds with the long term reason for doing it.

Let me discuss 1 – 4 together, each as part of one larger issue. Then I will talk about 5 separately. The reason for the distinction is this. While the analogy is not perfect, we can compare IT provider to faculty, on the one hand, with Department of Defense/Federal Government to citizenry, on the other. With that, 1 – 4 are meant to cover dealing with insiders, while 5 is meant to cover dealing with outsiders.

I think any IT leader’s motivation comes from wanting to get things done, things the leader himself values and would view as substantial accomplishments, accomplishments that if achieved would earn recognition for the leader. All these years later after some substantial experience both as an IT user and as an administrator, I confess that I have no good sense of how long things should take in implementing new IT projects, especially when they are done at scale. Let me start with that lack of as the base. To that let’s add the following. I know that in large organizations there is a problem with members who aren’t at the top feeling that they have to avoid blame, for which they’ll be singled out, in part because they won’t be granted the offsetting credit for accomplishment, which is either attributed to the organization as a whole or to the folks at the top. In football, the quarterback takes the offensive line out to dinner to partly address this issue. But the head coach doesn’t do that. And if the head coach is determined and not very communicative, he can make his players clam up and underperform. That is part of the issue.

The other part is the generational thing and the difficulty staying current with the technology. Those staff who are self-protecting in the job may not spend enough time learning new tricks, because the latter don’t have an obvious payoff. So they get stale in what they know and their inclinations and tendencies cause the leader to wince. The leader starts to wonder whether these people are sharp enough to do their job. Let’s get rid of the whole lot and start anew. Over time this becomes a grind where there is contempt on both sides. And the situation is exacerbated because in terms of financial incentives, there really aren’t any aside from promotion.

The need to get things done quickly comes from several sources – keeping up with peers, general developments in the world of IT that make change the norm, and a what-have-you- done-for-me-lately approach in how we are evaluated by our superiors and in how we evaluate ourselves. Basking in past glory might work when we’re talking about lines on a CV and publications from some years back that still get cited. And basking in past glory might create a habit of mind to do things right the first time out. But when in a service provider role, ancient history is just that and the present tense is the only one spoken. This is the same for the Department of Defense and us in learning technology. This living in the present tense puts enormous pressure on getting things done quickly.

The issue with cost is the eyes-bigger-than-the-stomach problem. We all want to implement the newest and greatest systems, because then all our users will clamor toward us and want to participate in what we provide. But if we cost out these systems fully in a sober manner, it’ll become clear that frequently we can’t afford them. So we do just what those subprime borrowers did in financing their homes that they no longer can continue making payments on. They literally mortgaged their future. So did we. Rumsfeld undoubtedly argued for the low troop numbers in Iraq after the invasion, because a higher assessment up front would have meant not doing it. This part of human nature is a bitch.

Among the various assessments of the Bush White House, one that seems strong is that early on Bush didn’t hear contrary argument to positions he ultimately advocated for. Nowadays he does hear about issues from more than one side, and his approach has been more conciliatory. One reason to insist you are right is to avoid the argument. Another is because the argument takes time to resolve. A third is because making the argument might give credence to those with whom you find fault. And yet another is that coming up with the idea to begin with is a creative act and creative acts should not be compromised by practical low-to-the-ground implementation considerations.

Rumsfeld was an innovator but he was guilty of all these sins. What about us in learning technology? How do we get change without making these mistakes?

Let me turn to the WMD issue. WMD was the smoking gun, the reason we had to invade Iraq, with no time to lose. (To be fair, WMD is more Cheney than Rumsfeld but for certain purposes they acted as a tag team.) At this time I still don’t know whether those guys really expect to find WMD or if they knew fully that it was a ruse, because while all evidence I’m aware of points to the ruse hypothesis, I can’t for the life of me understand how they could expect to pull it off. In other words, if from the point of view of 2002 they knew all the future until now, would they still have wanted to go through with it? Of course the number of troops issue matters here, so they could have believed Rummy was right on that even if WMD was a ruse. Even that is hard to believe.

In any event, WMD soon was abandoned for bringing democracy to the Middle East, perhaps a noble idea (but as anti-Reagan as one can imagine, Jeanne Kirkpatrick would have none of that). And of course the democracy idea has morphed since to something that might actually be sustainable. All of this (and Katrina too) led to a throw-the-Bushies-out mindset in the population as whole. Fortunately, we have the 22nd Amendment to the Constitution, which guarantees new leadership in the White House in January 2009.

There are no term limits for IT leadership (though the average tenure of a CIO is certainly less than 8 years. Maybe this is why.) And some of us will be long timers at our institutions for spousal, family, ties to the community, loyalty to colleagues, and other reasons. So why do we go off half cocked and make arguments akin to WMD about why such and such system will revolutionize learning on our campuses and we’ve got to have it or else? Have we thought through what it will be like two or three years into implementation of the system? Or does optimism reign eternal and we make these mistakes over and over again, because we are compelled to do so?

Let me close with this observation. I’ve exaggerated more than a bit in this piece. I’m not Donald Rumsfeld and I know that. Learning technologists and other IT leaders aren’t Donald Rumsfeld either. They can be sure of that. But they can spend some useful time scratching their heads about wherein lies the difference. That is something to ponder.

No comments: