Friday, November 30, 2007

"The median price for an existing home plunged by a record 5.1% in October"

That's the beginning of the 2nd sentence in a USA Today cover story from yesterday. Alarming, isn't it, that home prices would fall 5% in one month? But of course that is not the case. Home prices were 5% lower in October than they were A YEAR AGO-- big difference. I was suprised that a major newspaper would state something in such a misleading way, and I suspect that a lot of people don't notice the difference-- see my Rule about numeracy. Whenever I see charts tracking the real estate market, I wonder how many people confuse graphs that track the monthly percentage change in prices vs. the previous year with graphs that track the prices themselves. Either way the news isn't that good right now, but it's not as bad as USA Today made it sound yesterday!

9 comments:

Curtis Miller said...

Wow, that sounds just like a conversation I had with the MBA Quantitative Analysis class I'm teaching just this past week.

The answer to the question about whether you can make statistics say what you want is a resounding YES. You must be informed of how the test was made and what it is actually measuring in order to determine for yourself if the numbers being reported are meaningful.

MEG said...

The real estate stats are so often sensationalized and misrepresented that it really disgusts me. Even if/when they're technically true, they're taken out of context. Every time I see something about "record drops" in housing prices, I want to scream "That's because we had record increases last year!!!!" Real estate prices are still incredibly high (and even inflated) in many areas.

Imagine if the stock market shot up to 15,000 tomorrow for some reason and stayed there for a few months. Then it drops to 13,000 in 2008. All the stories would be talking about "THE WORST STOCK MARKET DECREASE SINCE THE GREAT DEPRESSION!!" But obviously we'd still be in great shape, and prices would be much higher than historical norms.

Mike said...

I don't think I can agree w/that comment about real estate prices dropping 5% in 1 month. They do all the time, every single year! When you look at the median sold price from month to month, they drop as you move from the spring to the fall to the winter. Then they spike up again as spring draws near. I would be MORE concerned over a 5% YOY (year-over-year) drop than a 5% monthly drop. That means the real estate market is TRULY dropping due to market conditions, and not due to seasonal fluctuations.

Anonymous said...

It's just another flashy headline that is supposed to attract a little buzz and attention. Yes it's misleading, but newspapers do it all the time.

Anonymous said...

Sensational headlines are all the rage these days, even if it means sacrificing the truth. Even the most reputable newspapers are doing this, more and more, I have observed.

Bitty said...

Yep, I came to say what limeade said and then some. Today's news organizations, including those that are supposed to be prestigious, are all about the sensationalized, misleading spin these days.

It's a national disgrace.

Anonymous said...

Well, you know the "first rule of journalism," right? "Never let the truth get in the way of a good story."

Andrew Stevens said...

Curtis, I just want to point out that the claim about whether you can make the statistics say what you want has nothing to do with statistics, per se. People can lie about whatever they like. The reason why statistics are easy to be deceptive about, without actually telling an untruth, is because most people don't understand them. Actually, I like it when people are misleading about statistics (which usually has more to do with ignorance and misunderstanding than with malice or deception). When they're misleading about statistics, they're much easier to correct.

I remember reading an article once where the headline was something like "Women Most Likely to be Murdered in First Year of Life." From the headline and opening paragraph, you would have thought it was showing that there was some sort of sexual selection going on in America with parents murdering their infant girls. In fact, infant boys are slightly more likely to be murdered than girls during the first year. The reason for the headline is that a woman's most dangerous age for homicide is the first year of life. They never again have a homicide rate that high. This is not the case for boys who are even more likely to be murdered between the ages of 18 and 24 than they are in their first year.

But as anonymous above said, the actual facts don't make for a catchy headline, since they accord with what people probably already know.

Anonymous said...

I blame the damn republicans!