An analysis of US soldier deaths in Iraq
A comment:
Here is quick analysis that I did of US soldier deaths in Iraq. I am a scientist, but I do not work with these types of data, so there may be flaws. However, since this is not a scientific publication and my results are interesting I decided to post them. Data, such as the kind that I used in this analysis, are readily available on the web. Given we trust the source, individuals who have some statistical ability can analyze these data and reveal information that may not be widely available in the mainstream media.
The Analysis:
Media attention has recently focused on a grim milestone: the total number of deaths of US soldiers in Iraq has topped 2000. While this figure is well worth mention, one aspect that is mostly missing from media coverage is that the rate at which these deaths occur may be increasing.
I obtained monthly data on US military deaths from the Iraq Coalition Casualty Count website [1] and fit a linear regression line to the data in order to extract an expected mean number of deaths per month. If the expected means show a rising trend then that can be interpreted as a rise in the death rate. For this analysis I removed March 2003 from the data because that month represents the first 12 days of the Iraq invasion (the deadliest 12 day period of the war). The invasion included hazards that were not present during the subsequent occupation, because the US military had not held ground before the invasion, and should not be compared with deaths from the rest of the occupation. Baghdad fell on April 9th 2003 so the occupation began after that date. Since the data I had available to me did not include daily figures I took a conservative approach and left the entire month of April in the dataset; doing so does not greatly influence the results.
What do the results say? The regression line I constructed explained 12% of the variation in the number of deaths. This means that the number of soldier deaths fluctuates greatly from month to month. Thus, trends are difficult to extract and should be interpreted with caution. Nevertheless, the expected mean for the beginning of the occupation in April 2003 was 47 deaths per month, while the expected mean for October 2005 was 80 deaths per month for a difference of 33 deaths per month. These numbers indicate a 70% increase in the death rate of soldiers from the beginning of the occupation to the present. If reflective of a real trend, this means that the occupation of Iraq is becoming far more deadly for American soldiers.
Here is quick analysis that I did of US soldier deaths in Iraq. I am a scientist, but I do not work with these types of data, so there may be flaws. However, since this is not a scientific publication and my results are interesting I decided to post them. Data, such as the kind that I used in this analysis, are readily available on the web. Given we trust the source, individuals who have some statistical ability can analyze these data and reveal information that may not be widely available in the mainstream media.
The Analysis:
Media attention has recently focused on a grim milestone: the total number of deaths of US soldiers in Iraq has topped 2000. While this figure is well worth mention, one aspect that is mostly missing from media coverage is that the rate at which these deaths occur may be increasing.
I obtained monthly data on US military deaths from the Iraq Coalition Casualty Count website [1] and fit a linear regression line to the data in order to extract an expected mean number of deaths per month. If the expected means show a rising trend then that can be interpreted as a rise in the death rate. For this analysis I removed March 2003 from the data because that month represents the first 12 days of the Iraq invasion (the deadliest 12 day period of the war). The invasion included hazards that were not present during the subsequent occupation, because the US military had not held ground before the invasion, and should not be compared with deaths from the rest of the occupation. Baghdad fell on April 9th 2003 so the occupation began after that date. Since the data I had available to me did not include daily figures I took a conservative approach and left the entire month of April in the dataset; doing so does not greatly influence the results.
What do the results say? The regression line I constructed explained 12% of the variation in the number of deaths. This means that the number of soldier deaths fluctuates greatly from month to month. Thus, trends are difficult to extract and should be interpreted with caution. Nevertheless, the expected mean for the beginning of the occupation in April 2003 was 47 deaths per month, while the expected mean for October 2005 was 80 deaths per month for a difference of 33 deaths per month. These numbers indicate a 70% increase in the death rate of soldiers from the beginning of the occupation to the present. If reflective of a real trend, this means that the occupation of Iraq is becoming far more deadly for American soldiers.
posted by Metatree at 3:41 PM
0 Comments:
Post a Comment
<< Home