Understanding Confidence Intervals

There are a lot of ways to analyze your survey results. One type of statistic you need to find is the confidence interval. For those unfamiliar with confidence intervals, they are a statistical representation of the “Confidence” you can have that the population’s actual responses fall within that specific range.

For example: – You run research and are looking for results with 95% confidence. – You get a score of X, and find the confidence interval (CI). – You now know with 95% confidence that the population’s actual number based on your sample is between X-CI and X+CI, You can consider it the margin of error, where you can say with 95% confidence (or more, depending on your analysis) that the population falls within that range. Yet people often forget what confidence intervals mean, and misconstrue the results. Here are several things to consider with confidence intervals. – All numbers within the interval are equally likely.

This is where a lot of researchers get confused. They see the confidence interval and assume that most likely the number falls in the middle of the range. Yet it’s equally possible that the actual number that represents the population is the lowest or highest number in the range – equally as likely as the number you got with your research.

Always remember to be conservative about your interpretation, so if the part of the margin of error shows that, say, an idea may not be worth pursuing, you may want to either run your research again or consider not pursuing the idea, even if the score you received from the sample would have originally indicated that it would be a good idea. – Confidence intervals are still affected by question biases. If you run a survey and your question has biases, then the confidence interval you receive may not be indicative of the feelings or beliefs of the population you’re trying to research – only how the population answers your bias question. – There is no 100% margin of error. It’s important to remember that 95% confidence intervals are just that – intervals with 95% confidence.

It’s still possible that actual number is part of that 5% that falls out of the interval. You may want to increase the amount of confidence you want (95% to 99% or more) or consider increasing the size of your sample in order to be more confident in your results. Confidence intervals are a very important part of you research.

They let you know how confident you can be that your results are meaningful. But confidence intervals are only valuable if you understand them. Be careful not to fall into common CI mistakes so that you can interpret your research correctly.