Tuesday, October 14, 2014

Everyone Knows

I once did analysis on the results of a community survey. When we took the top three issues as reported in aggregate, we found that nearly 100% of everyone would know someone with one suffering or facing one of these social issues. We related it this way: "Step outside your front door. Look at the houses to the left and right and across the street. One of those houses has someone involved in substance abuse, domestic violence or feeling underemployed and struggling financially." Mitigating this somewhat was the knowledge that perhaps the issues weren't as great as being reported. If everyone knows the town's only drunk then, when it's asked if you know anyone who is abusing alcohol, 100% of the respondents will say "yes". An analyst might think that 50% of the people are abusing as reported by the other 50% of the population or equal to whatever the responding percentage is.

Same problem can occur with employee surveys. If the questions are asked about what people know is happening around them, then the problem might appear worse than it is. "Do you know of any sexual harassment taking place in our company?" The percentage of affirmative responses might be inflated by the 'town drunk' syndrome. "Do you know of anyone who left the company because they were underpaid?" "Has anyone left because of a toxic boss?" "Is top management's communication clear to you and your co-workers?" (One person who shares their confusion over ambiguity, inaccuracies, etc. would taint the percentage of 'yes' responses.) "Are managers approachable if you have a work problem?" (One unapproachable boss that everyone knows...you get the idea...)

I'm not suggesting that employee surveys shouldn't be done. They should be done carefully and the questions should be phrased carefully. Particularly with a captive population when you're likely to receive a high percentage of responses out of the whole group, you don't need to ask if you know someone/anyone who has had a problem...Just ask each respondent if they have had a problem personally. Don't let them report on cafeteria conversations. It'll inflate the numbers.

No comments:

Post a Comment