The dawn of a new decade is upon us. Wow…seems like just yesterday I was getting ready for the non-event that was Y2K and now 2010 is here. For those of you keeping track at home, this also means that it’s time for good ol’ Uncle Sam to send around those two pound envelopes containing your U.S. Census form. As if being constitutionally obligated to fill out and return this wasn’t incentive enough, the Census Bureau is kicking off its “2010 Portrait of America Road Tour”, a $340 million publicity blitz that they are hoping will increase awareness and calm fears as to what the census population count really means. They are even planning on airing two commercials during the Super Bowl pregame show. My fingers are crossed that these ads have either Peyton Manning or monkeys in them. Everyone knows that either of these are comic gold when it comes to TV advertisements.
I, for one, think that this is a great idea (the Road Tour, not my commercial proposal). Taxpayer money has been tossed around carelessly for years; everything from the recent bank bailout to the continuous funding of space exploration. Yeah, we get it…the sun, the moon, the stars. All very nice. Show me video of a real live Ewok and I’ll be impressed. But I digress. My point is that there are several reasons why this type of initiative could payoff for data geeks like me in addition to resulting in better appropriation of government funds (I’ll be focusing more on the former).
Refresh of Stale Data Think back to the year 2000 and then think about how much has changed since then. Tiger Woods had a squeaky clean image, an iPod sounded like something an alien would pop out of, and when you thought of the term “hybrid” the image of a centaur or dog-faced boy would pop into your head before that of a car. So how can we rely on data that is 10 years old to provide us with accurate info for making decisions or performing statistical analysis and projections? How many housing developments have you seen spring up in your neighborhood and the surrounding area, especially during the real estate boom? Look back on some of the awful events of the last decade that have pushed people from their homes. All of these events need to be accounted for, and the more precise the reporting, the more useful the data. The fact that the census only happens once a decade makes this even more important. There isn’t a lot of room for error here because there won’t be another opportunity for quite some time (10 years to be exact).
Improved Sample Size There’s an old adage that says that one out of every 10 people is nuts. So think of nine people that you know and if none of them are crazy then you must be the one. But expand that to think of 11 people and maybe there’s a slight chance that you aren’t the one. Move on to 20 or 30 and all of the sudden you can justify your sanity a little bit easier. The same can be said with increasing the sample size of the census (ok, so my example was a bit of a reach). In a perfect world, 100% of the country would be represented by census data and we’d have an exact picture of the U.S. population, if only for a moment. That is completely unrealistic. More realistic would be an improvement over the mail return rate of 67% from 2000. As long as any gains shown represent a somewhat even distribution across the country then this would be seen as a positive. The closer this percentage gets to 100, the more confident a user can be in the data that they are consuming. I imagine that one concern would be that any gains in the response rate would come from specific pockets of the country where extra attention was paid, which could skew the overall results. However there are people way smarter than I am analyzing these types of things, so I don’t think there’s reason to panic.
New Data Available As an avid “Seinfeld” fan, I’ve seen pretty much every episode about a hundred times. In one episode in particular, George buys a sweater for Elaine that he knows has a spot on it because he can get it at a discount. This becomes a running gag throughout the show and hilarity ensues. If you think of the 2000 census as the sweater, then the 2010 census is taking several measures (or at least there are measures proposed) that would remove this spot. Things like tallying results from married same-sex couples, counting prisoners in their communities as opposed to where they are incarcerated (proposed), or increasing the representation of the Latino community would all go a long way towards painting a better picture of what is going on out there. A change in methodology would make trending a little bit more difficult but it would be laying the foundation for more accurate reporting in future censuses. Tradition isn’t an excuse for not making change, especially at the expense of accuracy. If the country’s current landscape warrants a change to the data that needs to be captured, then so be it.
I know the census isn’t a perfect system but at least efforts are being taken to bring about some positive changes. You can do your part by fulfilling your civic duty and sending in your forms after you receive them. It’ll be 10 years before you have to do it again and if movies about the future have taught me anything, it is that in 2020 everyone will have robot servants to fill out these forms for them. That and hover-cars.