... (or is that phalluses?). I about spit out my coffee when I clicked on a graphic from NuSI that Josh linked to in comments in another post. (BTW, an aside, I try to keep the blogging up at the expense of responding to comments so I apologize for getting quite behind of late.) So I decided to use my crudely constructed "arrows" at right in my graphics. Yes, that's Rick Moranis from Space Balls there!
Now one of the things that we're correctly lectured to about in the community is the shortcomings of observational studies: correlation does not equal causation. This is true. And yet, as I pointed out in Graphs, pictures showing correlation are frequently used to give the impression of causation. Now the graphic at right is not your typical correlation plot, but the message is clear. The reduction in supposedly "bad" fatty foods and increased consumption of carby foods, at least in part considered "good", has caused our obesity epidemic. This is more of the same-old same-old cherry picking of data to guide the consumers thought patterns to reach a desired conclusion.
Now NuSI and most low carbers like to cite the NHANES data for a relatively narrow time frame and/or convenient time points. I believe this is an innocent oversight to title the slide "food availability" and attribute the data to NHANES. But this series of who knows how many posts will address graphs either directly obtained from U.S. Per Capita Food Supply Trends, or constructed from data upon which this report was prepared from the USDA. Now there's no right or wrong data set, but I tend to believe supply data over self-report consumption data, given the plethora of data showing wide discrepancies between reported and actual intake. While we Americans can be a wasteful sort, the USDA has adjusted the supply data to include waste from 1970 forward. And for all the flack over food being big business, we also know that what doesn't get sold (and consumed) eventually doesn't get produced any more. Thus while actual consumption may not be determined with great accuracy, there's little reason to believe that they don't track to the availability/supply.
So ... I had a little fun!
Here's NuSI's 5 "food groups" using the 1970 - 2000 USDA Adjusted Per Capita Availability:
So now, animal protein is contributing to the obesity epidemic! Magic! But wait, there's more. I scoured the data for some general food groups with some rather unexpected changes, or changes NuSI might not find so compelling to put in their promotional graphics.
Note that citrus includes oranges and grapefruits, cream includes all kinds (half&half, light and heavy) that all saw increases despite a 24% decline in all milk consumption (IOW, folks weren't replacing cream in their coffee with low fat options), and green leafy veggies include lettuce, broccoli, spinach and squash (odd, but not my distinction).
Of course this doesn't tell us anything but ...