Aid Thoughts has an interesting post from yesterday on Hans Rosling and the presentation of data.
Rosling is a Swedish statistician who has done some great work on data visualization as it relates to traditional development issues (poverty, health care, education). He's also a sword swallower (starting at 18:30 in this video). But he's most famous for a couple of Ted talks, like the one below, where he shows changes in poverty levels and human development since the early 1800s.
The above video, btw, is the 7th most watched TED talk of all time.
What Aid Watch comments on, however, is another, more recent talk he gave in Doha (see below), where the razzle dazzle of the visuals seems to obscure the message. This is important, of course, because the message is what matters most for decision making. Focusing on what the data looks like, as opposed to what the data means, is a great way to get yourself in trouble. I have bad memories of when I was in Iraq with the Army and spent way too many hours putting together PowerPoint slides that would be looked at for 30 seconds, when that time would have been better spent trying to understand what the underlying causes behind the trends were, not to mention whether or not the data was reliable and the metric we were using was even valid as an indicator of success.
What also bothers Aid Watch are the "broad" conclusions Rosling draws from the data and his discounting of the role religion plays in country fertility rates. Rosling addresses the religion/fertility rate issue for two reasons, first, because most people believe there is a limit to how large a population the earth can handle is (hello, Malthus), and second, because some folks believe religion plays a decisive role in high fertility rates. Rosling seeks to destroy this latter argument and argues that we've reached "Peak Child," where as with "peak" anything, the trend either levels off or starts to decrease.
Rosling makes the case that fertility is more dependent upon variables such as child mortality rates, the need for child labor participation, women's education levels, and the accessibility of "family planning" services. Yet, the question not being looked at, according to AW, is whether or not religion plays a determinant factor with these variables. It's a good point. If it does, it then follows that religion plays a significant role in fertility rates. This is important to know because understanding causality impacts your decision making for where and how to allocate limited resources. Another development blog, A View From The Cave, also tackles the issue.
I bring this up because lately I've been doing a lot of research on data analysis in conflict stabilization and development situations and sorting out causality is an incredibly difficult task. If Rosling is right, then if you want to lower birth rates, you target the variables he mentions. But if he's wrong, you need to do something like trying to influence religious leaders who can sway the population. It makes sense, after all, to address the root cause of a problem as opposed to something in the middle. There is only so much time and money that practitioners have to address problems and sussing out true causality is particularly important when lives are at stake.
In Iraq and Afghanistan it seems that we haven't done a good job connecting our goals with our metrics for measuring progress, or making sure our data was both reliable and valid as indicators of success. Since the blog is back up, this is an issue I hope to cover a lot more in the future (in'shallah).