Ignoring reality and faking the stats

Ed at Dispatches from the Culture Wars posted about a False Comparison of Military Deaths. This was his response to a couple of blog posts that make claims completely failing common sense tests.

The first post is: Active Duty Deaths: Bush vs. Clinton, here’s a quote:

US Military active duty deaths simply have not gone up that much despite the fact that we’ve fought two major wars, liberating two formerly oppressed countries, and have struggled against an active terror insurgency in both countries ever since.

The second one is: US Lost More Soldiers Annually Under Clinton Than in Iraq, I won’t bother with a quote, the title clearly shows the premise. Ed does a good job pointing out the errors using the official D.O.D. statistics .

Wishing a more detailed analysis and wanting to get more practice with the greatly enhanced charting capabilities of Open Office I created a spreadsheet. I put the official data in a sheet and constructed some graphs to see if the data agreed with my common sense gut feeling that there are more military deaths in war time than peace time. Here is the result (click the graph to see the full size image).

mil_deaths_graph1

Looking at this chart there were clearly more deaths in administrations before and after Clinton. In fact the absolute lowest point is the final year of the Clinton administration.

Some commenter’s at the three blog posts thought that if you accounted for the differing sizes of the U.S. Military over time then GW Bush’s time would look better, one commenter even claimed that there was a 50% change in military personnel population. A 50% change in a population would certainly have a large impact on the graph so I investigated. Here’s a chart showing the change in U.S. Military population over time:

mil_deaths_graph3

Clearly the military population has changed over time however, it sure doesn’t look like 50%. In the spreadsheet I calculated the percent reduction from the high in 1986 and the low in 1999 and the drop was 35%. Even though the worst case drop is 35% I felt it could still have a major impact on the previous graph so, I did a variation were the death counts are expressed as the percentage of the total military population. This is the result (click the graph to see the full size image).

mil_deaths_graph2

Accounting for the changing population certainly made a difference in the graph. The pre-Clinton years don’t look so bad but, the post-Clinton years look the same or worse. Of course I doubt any of the people buying into the, all Democrats are bad and all Republicans are good, philosophy will adjust to reality. I still am amazed at how firmly peoples blinders can be set, on an engineering list there’s one guy who blames everything on the Democrats. He claims they are at fault for any and all problems in the U.S. even problems that occurred when they were out of power in all branches of government, sigh. Myself I’m a registered independent, I dislike most politicians of all the political parties because they so often fail to grasp reality. Note: My favorite description of reality:

“Reality is that which, when you stop believing in it, doesn’t go away.”

Philip K. Dick, How to Build a Universe That Doesn’t Fall Apart Two Days Later , 1978

You can download the spreadsheet and a PDF of the main spreadsheet data and graphs to review my work.

BTW – Open Office Calc is improved to the point were I think I can finally abandon MS Excel entirely.