Story by Rich Donnell,
I am sure you’ll notice in this issue our extensive coverage of our annual Sawmill Capital Expenditure Survey. The articles run from page 8 to page 34, so basically that is the magazine.
As usual, we’re providing two sections, one on the survey results of softwood lumber producers and the other on the feedback from hardwood lumber producers.
Surveys and polls are interesting animals. Whoever conducts them is usually pretty sure the results are an accurate reflection, though one can’t be totally certain.
Even the major political polling machines leave themselves a significant margin of error, sometimes as much as 3%. Do you know what a 3% margin of error means? (I recently had to explain it to a journalism class I teach at Auburn University; it opened their sleepy eyes.)
A 3% margin of error means that each percentage result may be off as much as 3%. Sounds simple enough, until you look closer. Say that the current poll has Obama at 52% and Romney at 48%. Looks like a pretty tight race doesn’t it? Well, a 3% margin of error means you can add and subtract 3% from each one of those figures at the same time. So 52-48 could really mean Obama 55%, Romney 45%. Not so close now. Or, on the other extreme, it could be that Romney is actually the one ahead, 51% to 49%.
So the numbers you see can lie, unless you read the fine print and understand it. Indeed a poll’s margin of error is usually shown in fine print adjacent the results. In addition, a margin of error, using our example, also means that there’s a 95% confidence that the poll is accurate including the 3% margin of error factor. So frankly, there may a little bit more give and take beyond the 3%.
Don’t ask me what our margin of error is in the softwood and hardwood lumber surveys. That is, don’t ask me if the survey result we show, for example, that says 71% of softwood lumber producers did not take more downtime in 2011 compared to 2010, and that says 29% did take at least as much downtime, could really mean 68-32 or 74-26 or something in between.
These surveys aren’t meant to be as “scientific” as the national political polls, but I believe they are accurate enough, as have been the surveys we’ve conducted in the past. We’re dealing with a very defined sampling, our industry readership, which we define even more into our readership e-mail list, to which we send the survey. We send the survey to this list twice. When we receive the results, we try and go through them and delete mill duplications and inconsistencies, and things that just don’t seem right based on our long-running association with the lumber industry. Of course it’s a relatively small percentage of producers who complete the survey, thus the margin of error, whatever it is.
One thing is for certain, the survey articles and charts make for excellent reading.