Wednesday, January 17, 2007

The Iraq Situation and Readily-Reusable Data

One of the difficulties I have encountered so far is that the data I would really like to examine first hand is not readily reusable and in some cases is completely unavailable in any form.

The first timeline collaboration principle that you will find on the right hand column of our blog template is:
1. Share the data series with the chart. Make sure it is readily reausable
To the best of my knowledge, not one of the 4 sources we have consulted so far (Brookings Iraq Index, The State Dept's Iraq Weekly Status, GlobalSeccurity, or the Iraq Casualty Coalition) has made the data that they used for their own analysis and charting both available and readily reusable as a spreadsheet or CSV file.

Admittedly some of these sources have shared their trend data series, but the method they selected means that they are not easy or ready for reuse without investing a lot of time and effort.

For example,
  • Brookings Iraq Index does include some trend data tables with important metrics and frequently labels the data values on their charts. Not all of the interesting trend data that appears in the tables is translated into a chart in the report. For example, see the table on page 5 of the January 8th report that shows details of the trends in cause of death for US Troops.
  • The Iraq Weekly Status Report from the State Department does not include any trend data at all beyond what they show in their charts. And the time period covered by the charts they have selected is rather short in most cases. For example the crude oil production chart only covers the last 10 weeks and other charts cover less than a year. Only the electricity chart stretches back to the beginning of 2004
  • Global Security includes a lot of detailed trend data in tabular form in addition to and complementing the charts they provide. The range of the data presented tabularly tends to cover the entire period from the start of the war until the present.
  • The Iraq Casualty Coalition also provides substantial amounts of tabular trend data in addition to their charts and their data covers the entire period from the beginning of the war.
Here's the problem.
Most people (both decision makers and ordinary citizens) are not particularly good or particularly thrilled by the idea of trying to make sense of trend data contained in a printed or online table.

My experience indicates that almost all adults (I would estimate the number at 90% or more) can make good sense of trend data presented in line graphs or bar graphs). On the other hand, I estimate that fewer than 10% of the people have the skills, interest or time to extract basic information from a trend table like the one on page 5 of the Brookings Iraq Index, or the many tables you can find at GlobalSecurity or at the Iraq Casualty Coalition.

What impact does this lack -- this lack of data / lack of readily reusable data / lack of easily viewable trend data for all key factors -- have on our ability to understand what's going on?

The way this plays out is that for 99% of the population, they will get their understanding of the trends from their most trusted source.

For a small percentage of people (I am guessing under 5% of the population), that trusted source might be one of the ones we have mentioned so far in this series on Iraq Trends or a similar detailed view of the situation that provides some depth and breadth and doesn't rely on just one or two metrics.

For this small group, if their source is Brookings, and they spend the time wading through the 30 page report, they will have a good chance of making good sense of the charts that Brookings selected. Similarly if their source is the State Department and if they wade through the Weekly Status report they will get a picture of the trends that the State Department selected but not the ones that Brookings was paying attention to. If their source is GlobalSecurity or the Iraq Casualty Coalition, they will get a good grasp of the charts presented at those sites.

Even in this small group that goes and gets their trend understanding from one of these useful and detailed sources (or other similar venues) a small percentage will dig any further into the tabular data that is available. The reason: primarily lack of time and secondarily lack of experience possibly combined with self conceived notions that reading tables is hard.

My reading is that few in this small group will consult and combine the trends from multiple sources like those mentioned above to get a broader perspective. Of course, if you have been reading this series of posts, we have done some of that lifting work for you and made it easier (still not easy in my opinion) to become more widely informed of the composite set of trends that are considered important drawn from a mix of analysts who may have vastly different axes to grind.

For a good portion of the rest of the population beyond this small group, the trusted source for trend data will turn out to be television and other mainstream media. Based on what I have seen from the MSM, this group of individuals will likely be exposed to quite limited trend information in any form and will often end up only being exposed to non-specific, non-numerical statements about the situation on the ground.


What to do? What to do?
Glad you asked. The answer for me is straightforward: ready reusability.

Ready reusability for the Iraq Trend data is easy to say and hard (time consuming, error prone) to do right now. In the next couple of posts, I will give you an example of what I mean by transforming some of the tabular data from expert sources into a readaily reusable form. Once the data is ready to go, I will provide some examples of why I think this is so important.

No comments: