Account for the Margin of Error

Think what you may, retailer, but a recent report of the public’s affection for cockroaches and the U.S. Congress has a relationship to assessing consumers’ approval of your store.
     News articles, such as on the Time magazine blog, are saying cockroaches get better approval ratings from American consumers than does the U.S. Congress. Those news reports are based on a press release from Public Policy Polling (PPP) about a survey in which 830 voters were asked “What do you have a higher opinion of: Congress or cockroaches?” Cockroaches prevailed with 45% of the respondents. Congress garnered 43%. The others said they were “not sure.”
     The Breitbart blog responded to the reports by saying the whole PPP project was rigged to make the Republicans in Congress look bad, with the objective of making President Obama look good by comparison.
     This Breitbart suspicion might be justified in the case of the cockroach survey question. You see, buried in the PPP press release is the statement “The margin of error is +/-3.4 percentage points.” This means that, in analyzing the survey results, a difference between percentages must be at least 3.4 in order for the results to be considered a true difference. This isn’t the case for the 45% to 43% comparison, so the most we can say is that the survey showed no evidence people rate the U.S. Congress as more worthy of approval than cockroaches.
     This isn’t good news for our federal legislature, but it’s not as bad as if the cockroaches won the poll.
     All reputable public opinion polling firms report the margin of error (MOE) for survey results. Unfortunately, MOE may not be included in customer satisfaction survey results reported to you by a consultant you hire, and that’s what all this has to do with retailing. If you hire a firm to survey customer satisfaction and you’ll be making important decisions based on the outcomes, be sure you’ll be told the margin or margins of error. Knowing this, you can interpret the survey results more accurately. (There are instances in which statement of a MOE would itself be misleading. In these instances, have the consultant explain why.)
     With surveys you do yourself for which you’ll use the results to make important decision, I recommend you have someone with statistical expertise show you how to calculate an MOE and teach you when and when not to calculate it.

Click below for more: 
Survey Consumers Person-to-Person 
Add “Please State” to “Other” Choice 
Increment Your Profitability Margins