The WSJ Online recently held an online survey to answer this burning question: "What proportion of your holiday gift spending did you do online this year?"
The results? As of December 26th:
- 31% of shoppers do 75% to 100% of their shopping online
- 24% of shoppers do 50% to 75% of their shopping online
- 15% of shoppers do 25% to 50% of their shopping online
- 30% of shoppers do less than 25% of their shopping online
So, according to the survey, over 70% of people do some shopping online for the holidays, and well over half do more than 50% of their shopping online! Given such strong numbers, we ought to go out and buy up all the internet retailers like Amazon and ebay, right?
Maybe. But first we should inspect the data. Is it truly useful? Is it even correct? We fully understand this isn't a scientific study and was never intended to be, but allow us to use this survey to illustrate a few points about the pitfalls of statistical measurement.
The study was conducted by the ONLINE Wall Street Journal. Which means you'd have to already be online in order to participate—automatically excluding anyone not already online. What's more, to take the survey you'd have to already be a subscriber to the online journal, which means you're already someone who makes online purchases. Call us sticklers, but asking people who already buy stuff online whether or not they buy stuff online is, um, a bit redundant and pretty skewed.
It's the equivalent of us asking whether you read online investment commentaries or not. (The answer would be 100%, since anyone reading this is by definition reading an investment commentary right now.) Again, we realize this particular WSJ survey was never intended to be scientific, but numbers like this get bandied about, cited, referenced and used as fodder for media pundits all the time. We'd even be willing to bet there are a few poor saps out there who'd use this information to invest with. Yet, ten seconds of scrutiny shows the data is worthless.
The point is that if you're going to really believe something enough to invest and stake your money on it, as often as possible you shouldn't trust other people's numbers. There's too much bad statistical methodology out there. Ask yourself how often you challenge convention or accepted wisdom. How often have you taken things to be true just because it's what everyone believes and always has?
This was just a simple online survey but believe us, you can find similar numbskull errors and bad assumptions in "official" statistics investors commonly use like GDP, employment, inflation, and others.
Successful investing is about being curious enough to be skeptical, ask questions, and find things out for yourself. Many don't have the time or ability to do this, which is why finding a trustworthy investment advisor who can do it on your behalf is all the more important. But whatever you do, ask those burning questions that others don't…or you're portfolio will end up scorched.