First, this is a somewhat goofy metric for stability. "80% of the data points within 20% of the mean" comes from an article by some heavyweight researchers in the field, but it is really just a way to see how closely data points cluster around the mean. Your data shows us an issue; the closer to zero your mean, the more minimal your stability envelope becomes. A mean of 20 provides for a data envelope of 4 points (+or - 2 points). A mean of 2 provides a data envelope of .4 points (+ or - .2). This makes mathematical sense, but in terms of human behavior, this is where things get goofy. If a client yells on average 20 times in an hour, that means 18 yells or 22 yells fits the envelope for stable data. But if they only yell on average 2 times an hour, then that means 1.8 or 2.2 times an hour fits the envelope. Hope that makes sense.
Hey, how to calculate my stability envelope if my mean is small (2.1) what would the envelope be?
First, this is a somewhat goofy metric for stability. "80% of the data points within 20% of the mean" comes from an article by some heavyweight researchers in the field, but it is really just a way to see how closely data points cluster around the mean. Your data shows us an issue; the closer to zero your mean, the more minimal your stability envelope becomes. A mean of 20 provides for a data envelope of 4 points (+or - 2 points). A mean of 2 provides a data envelope of .4 points (+ or - .2). This makes mathematical sense, but in terms of human behavior, this is where things get goofy. If a client yells on average 20 times in an hour, that means 18 yells or 22 yells fits the envelope for stable data. But if they only yell on average 2 times an hour, then that means 1.8 or 2.2 times an hour fits the envelope. Hope that makes sense.