0
$\begingroup$

I have a dataset of 540 observation like this:

facebook email age
1 0 23
1 1 34

1 a 0 are Yes/no, answers to the question: 'how do you find out about this survey?'

Can I measure correlation between various methods and Age? In order to prove that: smaller is the age, more people used facebook to find out the survey.

I tried to use a linear regression but it doesn't fit well.

I have to use logistic regression with age as dependent and others as independent?

$\endgroup$
6
  • $\begingroup$ What is your research question ? $\endgroup$ May 22, 2021 at 22:27
  • $\begingroup$ How do you find out about this survey? $\endgroup$
    – DOMEC
    May 23, 2021 at 8:11
  • $\begingroup$ I'm not sure what you mean ? "How do you find out about this survey?" you have 3 variables, so what research question are you trying to answer with this correlation coefficient? $\endgroup$ May 23, 2021 at 8:15
  • $\begingroup$ The correlation between age and different methods to finded out the survey. I found the distribution for age and answers, but i would like to know if it is possible to apply something as regression, to have more accurate results. Sorry,i Am not good in statistich $\endgroup$
    – DOMEC
    May 23, 2021 at 8:25
  • 1
    $\begingroup$ Regression could be a good choice, so which variable would your "outcome" be, and what which variables would be your "explanatory" variables ? $\endgroup$ May 23, 2021 at 8:28

1 Answer 1

1
$\begingroup$

I am going to challenge your question a little bit. When we do things like correlation or regression, it is usually because we have a sample of data and we want to use it to learn something about the whole population outside of the sample.

But in your case, the population outside of the sample would be "people who did not take the survey". Does it make sense to ask how people who didn't take the survey, found out about the survey?

If you are interested in the relationship between age and how they found out about the survey only for the people who took the survey, you have all the data already so you don't need to do correlation or regression. You can calculate the exact breakdown between groups for each age, or show the exact distribution of ages for each group - whichever makes more sense for your research.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.