My second semester college stats professor had this posted over his desk:
A statistician can have his head in an oven and his feet in ice, and he will say that on the average he feels fine.
Despite the fact that on multiple occasions I wanted to put either his head or mine in an oven, the guy really did teach me a great deal about the power, both good and bad, of statistics.
Now there is no way that six credit hours of statistics makes one an expert. Nor does sitting in meeting after meeting listening to semiconductor engineers drone on about F-tests and kurtosis. Been there, done that, made the charts. (Fair warning, don’t click those links without a bottle of aspirin handy.)
But having a little exposure to statistics does help one sift through the daunting pile of real estate stats that are published on a regular basis out there on the world wide web. Some real estate data analysis is very well done, some is not good but harmless in the grand scheme of things, and some is just flat out wrong.
Lately there seems to be a resurgence in the not good to bad category of real estate statistical analysis. I ran across a chart similar to this a couple of days ago:
What is this chart trying to tell us? The text said that it shows “units sold” in Phoenix, and that the upward trend was “happy news”. OK, I’ll buy that generally speaking, increased sales are a good thing.
But what does this chart really show us?
Absolutely nothing. Conspicuously missing is the Y-axis (vertical) scale. If you want to change the visual impact of a line chart, adjust the scale. Without a scale, we have no clue as to whether this chart means anything. Did sales go up by 10 units or 10,000? Isn’t that knowledge just a wee bit important if you are going to try to draw any conclusions from this data set?
Sadly, this chart (well, not THIS chart, but THIS one) wasn’t done by an individual who simply doesn’t know better. The chart I saw (but can’t copy without permission) was produced by REality, a company that claims to be “the leading provider of real estate market intelligence in North America” (an interesting statement given I’ve never heard of them — wonder if they have any data to support that claim?).
Business Week blows it too
The well-regarded publication Business Week provided an interesting example today of statistical analysis gone bad. In this brief article, BW tells us that traffic to certain markets on Realtor.com has increased and, “What this tells you is that a lot of people are starting to bottom-fish in these markets”.
Maybe that’s true, maybe not. To take one statistic — increased traffic — and conclude that means something in particular is quite the leap of faith. Traffic to this blog has increased dramatically this year, does that make it safe to say the Phoenix real estate market is booming?
What the BW article fails to explore, completely, are alternatives to why traffic may have increased on Realtor.com. Could it be due to a site redesign (which R.com went through), increased advertising in those markets by either local agents or R.com, a change in searcher demographics, a change in Google’s algorithms, or some weird shift in the space-time nexus? Who knows, all are plausible (well, maybe not the nexus thing, I pretty much made that one up).
“So what is the point Jay?” a bored and weary reader is likely saying right about now.
The point is, very few real estate agents (myself included) are well versed in statistics. Dare I say, neither are most main-stream media writers. Apparently neither is North America’s leading provider of real estate market intelligence. And you know what? The typical person looking for information about real estate is probably not all that familiar with statistical analysis either, though I have no statistics to prove that.
As a real estate agent building charts and tables filled with stats, keep in mind that conclusions you draw based on statistics may, or may not, be accurate. Do you really understand the difference between average, mean and median home price? Do your readers? Do you want to make predictions and draw conclusions from data you don’t really understand.
As a real estate “consumer”, be careful letting other people interpret data for you. You can’t believe everything you read on the Internet — unless it’s on this blog of course, he says dripping with sarcasm. In fact, it’s probably safe to assume the conclusion someone reaches from their “statistical analysis” is nothing more than their opinion that is loosely based on some data set. There is certainly nothing wrong with expressing an opinion, just be aware that’s what it is — an opinion. Do your due diligence, seek out varying opinions and analysis before reaching your own conclusions.
Hat tip to Ines at AgentGenius for the BW link.