The amount the percentage represents is irrelevant. A billion people could be involved, but if the total is 7 billion, it’s not going to be a significant part of the total trend.
5% can be a driver if it’s having a decent impact on your results. This is kind of a stats 101 thing man. You might even look for those outliers in your results and find a way to specifically exclude them if you find that the information you’re getting is being skewed. Do that too hard and it’s called P-hacking.
“We found that the bottom 5% of respondents were driving results negatively and so excluded the top and bottom 5%.”
Think about it as a literal driver. It’s a driver. It’s not the driver and also half the passengers. You can drive a motorcycle, you can drive a bus, and how much of the occupancy you are of those two things can change dramatically but you’re still a driver.
Less than 20% of a total is “significant”?
Yes. For example, 60 million people in the US (less than 20% of our total population) is a significant amount of people.
The amount the percentage represents is irrelevant. A billion people could be involved, but if the total is 7 billion, it’s not going to be a significant part of the total trend.
Yes it is…
5% can be a driver if it’s having a decent impact on your results. This is kind of a stats 101 thing man. You might even look for those outliers in your results and find a way to specifically exclude them if you find that the information you’re getting is being skewed. Do that too hard and it’s called P-hacking.
“We found that the bottom 5% of respondents were driving results negatively and so excluded the top and bottom 5%.”
Think about it as a literal driver. It’s a driver. It’s not the driver and also half the passengers. You can drive a motorcycle, you can drive a bus, and how much of the occupancy you are of those two things can change dramatically but you’re still a driver.