1
$\begingroup$

I have 2 Twitter datasets, one from 2014 and the other from 2015. I want compare the difference in Twitter volume (number of posts) between these two years. Would a t-test make sense here? I'm confused if I can compare volume instead of 2 means.

Thank you!

$\endgroup$
1
  • $\begingroup$ Please dexcribe the data and the question further. If you only have one count for one year and one count for the other, there is no way of telling, if one is significantly higher than the other, because there is no information, how large the counts usually vary from year to year. Further Explanation of the given data andrResearch question should lead to better answers. $\endgroup$
    – Bernhard
    Sep 2 '16 at 8:47
0
$\begingroup$

A possible idea is to apply a poisson glm model, so you can retrieve the significance of the categorical variable. E.g. on R:

y<-c(2164,2652,1236) x<-c("2001","2002","2001","2002") mod<-glm(y~x,family=poisson) summary(mod)

Also, you better dig a little bit more about the topic. There's plenty of publications or web pages, like https://www.researchgate.net/post/Which_is_the_most_appropriate_method_to_analyze_counts.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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