Page 1 of 1

How to calculate statistical significance of survey data

Posted: Sat Mar 25, 2017 12:25 am
by A7878
Hello,
I conducted a random survey about food labels of 132 people who were responsible for the grocery shopping in the household. My response rate was 80%. My population size is very large as it would be all the households in Canada (about 13,320 000) How best should I show the reliability of my data? How should I calculate my margin of error or confidence interval? I have calculated the mean of each question and I am putting the data in bar graphs and would like to put the margin of error on my graphs. I have looked at some calculations but I am not 100% sure of what I am doing. Thank you very much for your help.
Aiden

Re: How to calculate statistical significance of survey data

Posted: Thu Mar 30, 2017 3:26 pm
by LeungWilley
Hi Aiden,
This sounds like a great survey that you have conducted!

Unfortunately, I think we are going to need more details to be able to help. In very general terms, a lot of the question that you are asking would need to be designed into the survey. (For example, the confidence interval / reliability would be calculated with "checks" within the survey question set, i.e. asking the same question in different ways on the same survey and then comparing results. There are other ways of doing this as well.)

In regards to applying this data to the population of canada, you will need to "weight" the answer you received based on the Canadian population. (Demographic such as education level or income is one way to do this. It really depends on the factors (let say distance to grocery store, or how many time a week they go to the store) that gives you the most realistic picture as it relates to your experiment.)

Please post more information and we will try our best to help.
Good luck!
Willey