Reader survey 2015 – what you told us

More than at any other time in history, it is the responsibility of consumers, voters and business owners to study the art of the survey. The reason should be obvious. More than ever, consumers, voters and business owners are being asked to rely on poll results in favour of responsible governance and in favour of facts. It is a rare internet purchase, for example, that does not come with its own demand that you spend your time as well as your money to report on that product.

And, in too many cases, the truth of your opinion is buried under a pile of paid survey results, if negative, an elevated along with paid survey results if it is not. Properly done, a survey can provide leaders, producers and administrators with valuable information on the beliefs of their constituents and markets. Improperly done, they are transparent and fraudulent propaganda. We have all heard of a biased survey. What, exactly, is that?

Testing the variables

In statistics, the first variable to test is whether you have a valid sample, since the art of the survey is to discover the opinion of an entire statistical universe by only asking a percentage. The goal, then, is to get a representative sample. Typically, this is done by taking every Nth record across a valid section of the universe. If you want a 10-per cent sample, you would take every 10th name, being careful not to concentrate in only a few population centres or in only a few demographic fields. For example, you would likely not find a valid sample of cross-country skiing appeal by polling rural Louisiana.

The example, of course, is ridiculous, but it illustrates how a sample can be biased. I have always argued that all surveys are automatically biased by getting only respondents that respond. Not trying to be funny. In every survey, the majority of survey recipients do not respond, leading to the very important “response rate” statistic. And, while we use the respondents as the data base, every surveyor would love, just one time, to see what the silent majority would say.

How our survey works

Our survey is biased by another variable, as well. We send the survey only to people that have given us permission to use their e-mail for, as we have promised, only editorial-related issues. Typically, this means only the monthly e-letter. But twice a year – once for us and once for a research associate in the States, we ask you to complete surveys. However, while this biases the sample toward people who use e-mail and provide it, we feel that in 2014 our readers virtually all have e-mail and the ones that give us permission to use it comprise also the ones that would be likely to respond.

Therefore, while surveying is an inexact science, it can be controlled. And you can use this information to help you understand the purpose and the value of other surveys you are asked to complete.

Questions

Another bias is the bias of the questions. Despite what others may argue, we believe that every question is biased or it is not a question. For example, when we ask what you think of Canada’s floorcovering distribution system, it may dispose readers to think we want an answer of one type or another. But there is no way to ask the question and get a clear answer if you don’t ask a clear question. As you will soon see, Canadian floorcovering professionals are quite happy with the distribution system – a useful piece of information when looking at the sector as a whole.

Therefore, we have described our methodology. We use a random sample, taken across all geographic and population ranges, and try in the space of 12 questions or so to get a snapshot of the industry’s attitude. We are not trying to discover your buying intentions. That is another type of survey.

What you told us

Our first set of questions was designed to reveal the degree to which regulatory expenses have an impact on small businesses. The opening question was whether you have ever fired an employee for non-performance. You responded that 61 per cent of you had, and 39 per cent had not.

Of those who had fired an employee for non-performance, 34 per cent said they experienced regulatory costs over and above accrued vacation and benefits, and 66 per cent did not.

Of those who had experienced excess costs, 56 per cent said those costs were foreseen and had accounted for the expense in their termination decision. Meanwhile, 44 per cent had not foreseen the excess costs. The average indirect costs experienced by our respondents for firing a non-performing employee was $15,000.

In addition, of all employers who had fired a non-performing employee, 42 per cent consulted a lawyer before taking action. Another 58 per cent did not.  Still another 31 per cent consulted a lawyer after taking action and 69 per cent did not.

Respondent demographics

One of the interesting question series involved people’s vocation, i.e., we asked whether the respondent is a retailer or an installer. Of the respondents, 36 per cent said they are retailers; 64 per cent said they are not. Meanwhile, 18 per cent said they are installers; 82 per cent said they are not. Even if you assume all the retailers are also installers, this only accounts for 44 per cent of the respondents. We also include a large number of interested architects and designers, but not enough to account for the expected percentages of around 70 per cent, combined.

Possibly we can sort this out in a future study.