System Dynamics Conf. – II – Day 0 (PhD Colloquium)

Sunday was the PhD colloquium. It was smaller than the full conference has been, but there were still about 50-100 people in attendance. In the morning, there were several presentations, 3 of which I unfortunately missed due to my late arrival the night before, followed by a luncheon. Then, in the afternoon, there was a poster session, some meetings, and a really interesting panel.

After picking up my copy of the conference schedule from registration, one thing became really obvious – there’s a lot of stuff crammed into this conference. Six parallel paper tracks, as well as workshops, chapter meetings, special interest groups, workshops, and all the normal social events.  This is both good and bad – good because it’s rare that there’s nothing interesting going on, bad because conflicts between things you’re interested in are common.

Here’s the stuff that grabbed my interest on Sunday:

• Tommy Ka Kit Ngai from the U of Cambridge talked about the distribution and uptake of clean water filters in a South Indian case study. Interestingly, he sought to integrate the widely diverse ways that uptake has been explained in the past into a single model. There were two interesting points comments that came up in the discussion:
1. One guy was very keen to put a dollar figure on the cost of providing the filters to people versus the cost of not having clean water. Given the severely negative impact of tainted water on one’s health, this seemed kinda abhorrent, but his argument was that this would actually make it easier to convince governments and other NGOs to fund water filter distribution. Though it seems plainly obvious that clean water is a Good Thing, I guess he had a point in that the economic argument, ignoring all the humanistic and qualitative factors as it does, is more concrete and thus more compelling to bureaucrats.
2. Someone else speculated that randomness played a role in concealing the individual benefits of using a water filters. Imagine people get sick 4 +/- 2 times a year if they’re drinking dirty water, and only 2 +/- 1 if they’re drinking clean water. The difference is visible statistically, but in many cases, an individual’s sample size of one could make this difference quite difficult to discern. In cases like this, people are then reliant on the authority of outside agencies to tell them that water filters are good for them. Much better is to have a way of making this difference visible to them. One suggestion might be to have a glass jar sitting somewhere public in the community. Every day someone’s sick, add a pebble to the jar. Do this for three months, then give people water filters and do it again for another three months. Compare the jars. There’s obvious ethical and sampling problems with this, but you can see how useful the general idea of making the causal link popularly visible is.
• Met a guy at the luncheon with a truly magnificent mullet. He also gained bonus points as his reason for being at the conference was “I come here on vacation”. I heartily approve of that attitude.
• The ethics panel was really good. Model building is at the centre of many contentious issues in policy and business; salient examples include financial models and the current economic crisis, and climate models predicting climate change and destabilization. Furthermore, models often insulate decision makers from the systems they govern, and consequently model builders have a great deal of power and thus a great deal of responsibility. This means that ethics are critical in this discipline. The panel had three members: Alan Graham, Dennis Meadows, and Kim Warren. Here’s what I got from it:
  • Dennis gave a neat demonstration of the principle of “it’s not what you say, it’s what you do”. He told us he would count down from three, then say “Clap!”, at which point we were to clap in unison. He counted down, clapped himself, then, after a pause said “Clap!”. We, of course, all clapped when he did, not when he told us to.
  • Dennis also argued that professional ethics flow from personal ethics and that the best way to develop strong professional ethics is to focus on strong personal ones. He gave a lot of generally good advice about honesty, scrupulously honouring one’s commitments, knowing one’s limits, relating well with and supporting others, and general humility. Three cool bits: firstly, a quote by Janice Joplin “Don’t compromise yourself, it’s the only thing you’ve got”, which I quite liked; secondly, the idea that one is committing an injustice when one gets into relationships that benefit only one party because exploitation is wrong and charity only works when it’s genuine and not onerous; and thirdly, the suggestion that one should decide where one’s ethical boundary lines lie in advance of them being challenged, because otherwise it’s far too easy to make compromises that one is unhappy with.
  • Alan talked a lot about being honest concerning the limits of one’s tools. Like any tool, System Dynamics is sometimes helpful, but comes with built-in limitations. His recommendations: do not oversell the tool, and be careful not to let your clients overrate its effectiveness; be very careful not to mislead people or allow the perception that this may have occured; and, avoid even the smallest falsehoods, as these will be mercilessly exploited the moment your work becomes contentious. A couple of interesting points came out in his discussion. What do you do when the client takes your model and misleads others with it? To what extent should you get into technical details which might confuse or mislead clients? He suggested that one should clarify carefully the nature of one’s obligation to  client and, where possible, the uses to which one’s model will be put.
  • Kim built on this by talking more about ethical dilemmas. He outlined situations where ethical obligations to the client conflict with ethical obligations to society, then talked a bit about ways of rationalizing and making decisions in such situations. Like Dennis, he emphasized the importance of defining one’s ethical boundary in advance. His discussion also emphasized the need to think about the long term or indirect consequences of a model. While he didn’t seek to convince us that his that his approach was definitely right, he did seem to employ a fairly utilitarian mindset. So, for example, if one’s model suggests implies that a division of company is dragging that company down, it may be ethical to recommend the elimination of that division, despite the effects on the individuals working withing, if not doing so would risk bringing down the whole company, thus jeopardizing the jobs of everyone in the company. Another example involved offering strategy to a company such that they might act anti-competitively or even destroy a competitor. In that case, he said that after consideration, he resolved his ethical issues by making a value judgement and determining that the competitor being destroyed were harmful in some way, and that their elimination would be good for society as a whole. This is interesting, because it really reveals the subjective nature of ethics of this sort.