I am going through a course on linear regression and they say to check for independent observations: "based on the data collection process would the value of one data point impact the value of another data point". My question is, for a time series of text message sentiment between two people, is this independent? Because the values of one message would impact the following one because it is two people messaging each other.

  • $\begingroup$ Independence is a very specific mathematical characterization of a probability model. The check you describe is only one special way to determine whether it would be reasonable to assume independence. But whether a particular time series is well described by some form of independence comes down to the series itself: you have to inspect it to assess how consistent it is with your assumptions. $\endgroup$
    – whuber
    Commented Apr 3 at 15:24
  • $\begingroup$ How do I "inspect it to assess how consistent it is with your assumptions" $\endgroup$ Commented Apr 3 at 17:25
  • $\begingroup$ There are many, many ways: search our site for test independence for some common settings and procedures. $\endgroup$
    – whuber
    Commented Apr 3 at 17:27


Your Answer

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

Browse other questions tagged or ask your own question.